CN109963174B - Flow related index estimation method and device and computer readable storage medium - Google Patents
Flow related index estimation method and device and computer readable storage medium Download PDFInfo
- Publication number
- CN109963174B CN109963174B CN201910082401.9A CN201910082401A CN109963174B CN 109963174 B CN109963174 B CN 109963174B CN 201910082401 A CN201910082401 A CN 201910082401A CN 109963174 B CN109963174 B CN 109963174B
- Authority
- CN
- China
- Prior art keywords
- video
- flow related
- estimated
- information
- related index
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 238000010606 normalization Methods 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012216 screening Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 6
- 239000003550 marker Substances 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 3
- 238000013486 operation strategy Methods 0.000 description 12
- 238000004519 manufacturing process Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 7
- 230000000717 retained effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 230000009193 crawling Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/258—Client or end-user data management, e.g. managing client capabilities, user preferences or demographics, processing of multiple end-users preferences to derive collaborative data
- H04N21/25866—Management of end-user data
- H04N21/25891—Management of end-user data being end-user preferences
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/262—Content or additional data distribution scheduling, e.g. sending additional data at off-peak times, updating software modules, calculating the carousel transmission frequency, delaying a video stream transmission, generating play-lists
- H04N21/26208—Content or additional data distribution scheduling, e.g. sending additional data at off-peak times, updating software modules, calculating the carousel transmission frequency, delaying a video stream transmission, generating play-lists the scheduling operation being performed under constraints
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/81—Monomedia components thereof
- H04N21/8126—Monomedia components thereof involving additional data, e.g. news, sports, stocks, weather forecasts
- H04N21/8133—Monomedia components thereof involving additional data, e.g. news, sports, stocks, weather forecasts specifically related to the content, e.g. biography of the actors in a movie, detailed information about an article seen in a video program
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Graphics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a flow related index estimation method, a flow related index estimation device and a computer readable storage medium, and belongs to the technical field of computers. The method can determine candidate reference videos with similarity meeting a first preset condition with the to-be-estimated videos according to the related information of the to-be-estimated videos and the related information of each candidate reference video to obtain target videos, and determine estimated flow related indexes corresponding to different moments after the to-be-estimated videos are released based on flow related indexes corresponding to different moments after the to-be-estimated videos are released. According to the embodiment of the invention, the flow related indexes corresponding to different moments after the video to be predicted is released can be predicted according to the video similar to the video to be predicted, so that the flow related indexes of the video to be predicted can be predicted more accurately.
Description
Technical Field
The invention belongs to the technical field of computers, and particularly relates to a flow related index estimation method and device and a computer readable storage medium.
Background
In order to improve the flow of the video platform, the video platform side often needs to estimate the flow related indexes of different videos, such as the video playing times, the video playing duration, the number of watched users, and the like, and then makes different operation strategies for the different videos according to the estimated flow related indexes, for example, for a video with a large estimated flow related index, the popularization strength of the video is increased, and for a video with a small estimated flow related index, the popularization strength of the video is reduced.
In the prior art, when a traffic related indicator of a video to be estimated is obtained, related information of the video to be estimated, such as a director, a producer, and the like, is usually obtained, and then the traffic related indicator of the video to be estimated is calculated according to traffic related indicator values corresponding to specific objects represented by different preset related information and weights of different related information. For example, assuming that the specific object represented by the director is director a and the specific object represented by the offerer is offerer b in the related information of the video to be estimated, a product of a preset flow related index value corresponding to the director a and a preset weight of the director, a product of a preset flow related index value corresponding to the offerer b and a preset weight of the offerer b may be calculated, and finally, a sum of the two products is used as the estimated flow related index of the video to be estimated.
However, the estimation method in the prior art can only estimate a fixed flow related index according to the inherent related information of the video to be estimated, but is affected by uncertainty factors, such as the degree of public praise of the video, the transmission strength of the video by the user, and the like, and actual playing conditions of the video at different stages after being released may be different, so that the estimation accuracy in the prior art is low.
Disclosure of Invention
In view of this, the present invention provides a method and an apparatus for estimating a traffic related indicator, and a computer readable storage medium, which solve the problem of low estimation accuracy to a certain extent.
According to a first aspect of the present invention, a method for estimating a flow related indicator is provided, where the method includes:
determining candidate reference videos with the similarity to the video to be estimated meeting a first preset condition according to the relevant information of the video to be estimated and the relevant information of each candidate reference video to obtain a target video; the candidate reference video is a video meeting a second preset condition at the release moment;
determining estimated flow related indexes corresponding to different moments of the video to be estimated after release based on the flow related indexes corresponding to the different moments of each target video after release; and the different moments after the release represent time points which take the release moment as a starting point and are different from the release moment in duration.
Optionally, the determining, according to the related information of the video to be estimated and the related information of each candidate reference video, a candidate reference video whose similarity to the video to be estimated satisfies a first preset condition to obtain a target video includes:
respectively generating the characteristics of the video to be estimated and the characteristics of each candidate reference video based on the relevant information of the video to be estimated and the relevant information of each candidate reference video;
respectively calculating the Euclidean distance between each candidate reference video and the video to be estimated according to the characteristics of each candidate reference video and the characteristics of the video to be estimated to serve as the similarity;
and taking the candidate reference video with the corresponding Euclidean distance smaller than a preset distance threshold value as the target video.
Optionally, the related information includes different types of information;
before determining the candidate reference video with the similarity to the video to be estimated meeting the first preset condition according to the relevant information of the video to be estimated and the relevant information of each candidate reference video, the method further comprises the following steps:
respectively taking the video to be estimated and each candidate reference video as a video to be processed, and acquiring information corresponding to the video to be processed from different data sources according to the type of the information contained in the related information to obtain multiple spare related information corresponding to the video to be processed;
carrying out duplicate removal processing on an information set formed by the multiple spare related information to obtain a target set;
and screening the information included in the target set according to the specified confidence degrees of different data sources, and taking the information included in the screened target set as the related information of the video to be processed.
Optionally, the generating the features of the candidate reference video and the features of the candidate reference video respectively based on the related information of the to-be-estimated video and the related information of the candidate reference video includes:
for each type of information in the related information of each video to be processed, the following operations are respectively executed:
if the number of the selectable objects corresponding to the type of the information is not larger than a preset threshold value, setting the value of an element corresponding to the selectable object matched with the object represented by the information in the designated element sequence as a first mark value; setting values of other elements in the specified element sequence as second marker values; taking the set designated element sequence as a characteristic corresponding to the information; combining the characteristics corresponding to each type of information to obtain the characteristics corresponding to the video to be processed; the selectable objects corresponding to the type of the information correspond to the elements contained in the specified element sequence one by one; or,
and if the number of the selectable objects corresponding to the type of the information is larger than the preset threshold value, generating the characteristics of the video to be processed according to the network index of the object represented by the information.
Optionally, the determining estimated traffic related indexes corresponding to the to-be-estimated video at different moments after the release based on the traffic related indexes corresponding to each target video at different moments after the release includes:
if the video to be estimated is a published video, determining a predicted flow related index according to an actual flow related index corresponding to the video to be estimated and flow related indexes corresponding to different moments of each target video after the video to be estimated is published; or,
if the video to be estimated is a video which is not released, calculating average flow related indexes corresponding to different moments after releasing according to the flow related indexes corresponding to different moments after releasing of each target video; and taking the average flow related indexes corresponding to different moments after the release as estimated flow related indexes corresponding to different moments after the release of the video to be estimated.
Optionally, the determining a predicted flow related index according to the actual flow related index corresponding to the video to be estimated and the flow related index corresponding to each target video at different time after the target video is released includes:
for each target video, according to the corresponding flow related indexes of the target video at different moments after the target video is released, constructing a time and flow related index curve of the target video in a coordinate system formed by a first axis and a second axis; the first axis represents different moments after release, and the second axis represents flow related indicators;
according to the key moment corresponding to the target video and the flow related index corresponding to the key moment, carrying out normalization processing on the time and flow related index curve; the time interval between the key moment and the release moment of the target video is equal to the key duration;
acquiring an average curve of the time and flow related indexes according to the normalized time and flow related index curve of each target video;
and calculating the predicted flow related index according to the actual flow related index corresponding to the video to be estimated, the published time length corresponding to the actual flow related index, the average curve of the time and flow related indexes and the key time length.
Optionally, the normalizing the time-flow related index curve according to the key time corresponding to the target video and the flow related index corresponding to the key time includes:
and for each coordinate point in the curve of the time-flow related index, setting a numerical value corresponding to a first axis in the coordinate points as a quotient of the numerical value and the key duration, and setting a numerical value corresponding to a second axis in the coordinate points as a quotient of the numerical value and the flow related index corresponding to the key moment.
Optionally, the published duration is a time interval between a moment corresponding to the actual flow related index and the publishing moment of the video to be estimated;
the calculating the predicted flow related index according to the actual flow related index corresponding to the video to be estimated, the published time length corresponding to the actual flow related index, the average curve of the time and flow related indexes and the key time length comprises:
calculating the ratio of the published time length to the key time length corresponding to the video to be estimated to obtain a first ratio;
searching the flow related indexes corresponding to the first ratio from the average curve of the time and flow related indexes, and searching the flow related indexes corresponding to each predictable value; the predictable value represents a value in the first axis that is greater than the first ratio;
respectively calculating the ratio of the flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio to obtain a second ratio corresponding to each predictable value;
and for a second ratio corresponding to each predictable value, calculating the product of the actual flow related index and the second ratio to obtain a predicted flow related index corresponding to the moment of the video to be predicted after the release represented by the predictable value.
Optionally, after determining the estimated traffic related indicators corresponding to the to-be-estimated videos at different moments after the release based on the traffic related indicators corresponding to each target video at different moments after the release, the method further includes:
and determining alternative intervals of the flow related indexes corresponding to different moments of the video to be estimated after the video is published based on the flow related indexes corresponding to different moments of each target video after the video is published.
Optionally, the determining, based on the traffic related indicators corresponding to different moments after the release of each target video, the alternative intervals of the traffic related indicators corresponding to different moments after the release of the to-be-estimated video includes:
if the video to be estimated is a published video, determining alternative intervals of the flow related indexes corresponding to different moments after the video to be estimated is published according to the actual flow related indexes corresponding to the video to be estimated and the flow related indexes corresponding to different moments after the video to be estimated is published;
if the video to be estimated is a video which is not released, acquiring maximum flow related indexes and minimum flow related indexes corresponding to different moments after releasing according to flow related indexes corresponding to different moments after releasing of each target video; and regarding the maximum flow rate related index and the minimum flow rate related index corresponding to different moments after the release, taking the maximum flow rate related index and the minimum flow rate related index as end values of a section corresponding to the moments, and obtaining a candidate section corresponding to the moments.
Optionally, the determining, according to the actual traffic related index corresponding to the video to be estimated and the traffic related index corresponding to each target video at different time after being released, the alternative intervals of the traffic related indexes corresponding to the video to be estimated at different time after being released includes:
searching a maximum flow related index and a minimum flow related index corresponding to the first ratio according to the normalized time and flow related index curve of each target video, and searching a maximum flow related index and a minimum flow related index corresponding to each predictable value;
respectively calculating a third ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio, and a fourth ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio;
and for each predictable value, taking the product of the actual flow related index and the third ratio corresponding to the predictable value and the product of the actual flow related index and the fourth ratio corresponding to the predictable value as end values of an interval, and obtaining an alternative interval corresponding to the moment after the release represented by the predictable value.
According to a second aspect of the present invention, there is provided an apparatus for estimating a flow related indicator, the apparatus may include:
the first determination module is used for determining candidate reference videos of which the similarity with the videos to be estimated meets a first preset condition according to the relevant information of the videos to be estimated and the relevant information of each candidate reference video to obtain target videos; the candidate reference video is a video meeting a second preset condition at the release moment;
the second determination module is used for determining estimated flow related indexes corresponding to different moments of the video to be estimated after the video is published based on the flow related indexes corresponding to different moments of each target video after the video is published; and the different moments after the release represent time points which take the release moment as a starting point and are different from the release moment in duration.
Optionally, the first determining module includes:
the generation submodule is used for respectively generating the characteristics of the video to be estimated and the characteristics of each candidate reference video based on the relevant information of the video to be estimated and the relevant information of each candidate reference video;
the first calculation submodule is used for respectively calculating the Euclidean distance between each candidate reference video and the video to be estimated according to the characteristics of each candidate reference video and the characteristics of the video to be estimated, and the Euclidean distance is used as the similarity;
and the first determining submodule is used for taking the candidate reference video with the corresponding Euclidean distance smaller than a preset distance threshold value as the target video.
Optionally, the related information includes different types of information;
the device further comprises:
the acquisition module is used for respectively taking the video to be estimated and each candidate reference video as a video to be processed, and acquiring information corresponding to the video to be processed from different data sources according to the type of the information contained in the related information to obtain multiple spare related information corresponding to the video to be processed;
the duplication removing module is used for carrying out duplication removing processing on an information set formed by the multiple spare related information to obtain a target set;
and the screening module is used for screening the information included in the target set according to the specified confidence degrees of different data sources, and taking the information included in the screened target set as the related information of the video to be processed.
Optionally, the generating sub-module is configured to:
for each type of information in the related information of each video to be processed, the following operations are respectively executed:
if the number of the selectable objects corresponding to the type of the information is not larger than a preset threshold value, setting the value of an element corresponding to the selectable object matched with the object represented by the information in the designated element sequence as a first mark value; setting values of other elements in the specified element sequence as second marker values; taking the set designated element sequence as a characteristic corresponding to the information; combining the characteristics corresponding to each type of information to obtain the characteristics corresponding to the video to be processed; the selectable objects corresponding to the type of the information correspond to the elements contained in the specified element sequence one by one; or,
and if the number of the selectable objects corresponding to the type of the information is larger than the preset threshold value, generating the characteristics of the video to be processed according to the network index of the object represented by the information.
Optionally, the second determining module includes:
the second determining submodule is used for determining a predicted flow related index according to an actual flow related index corresponding to the video to be estimated and flow related indexes corresponding to different moments after the target video is released if the video to be estimated is a released video; or,
the second calculation submodule is used for calculating the average flow related indexes corresponding to different moments after the video is published according to the flow related indexes corresponding to different moments after the target video is published if the video to be estimated is a video which is not published; and the third determining submodule is used for taking the average flow related indexes corresponding to different moments after the release as the estimated flow related indexes corresponding to different moments after the release of the video to be estimated.
Optionally, the second determining sub-module includes:
the construction unit is used for constructing a time and flow related index curve of each target video in a coordinate system consisting of a first axis and a second axis according to flow related indexes corresponding to different moments after the target video is published; the first axis represents different moments after release, and the second axis represents flow related indicators;
the normalization unit is used for performing normalization processing on the time and flow related index curve according to a key moment corresponding to the target video and a flow related index corresponding to the key moment; the time interval between the key moment and the release moment of the target video is equal to the key duration;
the acquisition unit is used for acquiring an average curve of the time and flow related indexes according to the normalized time and flow related index curve of each target video;
and the calculating unit is used for calculating the predicted flow related index according to the actual flow related index corresponding to the video to be estimated, the published time length corresponding to the actual flow related index, the average curve of the time and flow related indexes and the key time length.
Optionally, the normalization unit is configured to:
and for each coordinate point in the curve of the time-flow related index, setting a numerical value corresponding to a first axis in the coordinate points as a quotient of the numerical value and the key duration, and setting a numerical value corresponding to a second axis in the coordinate points as a quotient of the numerical value and the flow related index corresponding to the key moment.
Optionally, the published duration is a time interval between a moment corresponding to the actual flow related index and the publishing moment of the video to be estimated;
the computing unit is configured to:
calculating the ratio of the published time length to the key time length corresponding to the video to be estimated to obtain a first ratio;
searching the flow related indexes corresponding to the first ratio from the average curve of the time and flow related indexes, and searching the flow related indexes corresponding to each predictable value; the predictable value represents a value in the first axis that is greater than the first ratio;
respectively calculating the ratio of the flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio to obtain a second ratio corresponding to each predictable value;
and for a second ratio corresponding to each predictable value, calculating the product of the actual flow related index and the second ratio to obtain a predicted flow related index corresponding to the moment of the video to be predicted after the release represented by the predictable value.
Optionally, the apparatus further comprises:
and the third determining module is used for determining alternative intervals of the flow related indexes corresponding to different moments of the video to be estimated after the video is published based on the flow related indexes corresponding to different moments of each target video after the video is published.
Optionally, the third determining module includes:
a fourth determining submodule, configured to determine, if the video to be predicted is a published video, alternative intervals of the traffic related indexes corresponding to different moments after the release of the video to be predicted according to an actual traffic related index corresponding to the video to be predicted and traffic related indexes corresponding to different moments after the release of each target video;
the obtaining submodule is used for obtaining maximum flow related indexes and minimum flow related indexes corresponding to different moments after the target videos are published according to the flow related indexes corresponding to the different moments after the target videos are published if the videos to be estimated are unpublished videos; and a fifth determining submodule, configured to, for a maximum traffic related indicator and a minimum traffic related indicator corresponding to different times after the release, obtain an alternative interval corresponding to the time by using the maximum traffic related indicator and the minimum traffic related indicator as end values of an interval corresponding to the time.
Optionally, the fourth determining sub-module is configured to:
searching a maximum flow related index and a minimum flow related index corresponding to the first ratio according to the normalized time and flow related index curve of each target video, and searching a maximum flow related index and a minimum flow related index corresponding to each predictable value;
respectively calculating a third ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio, and a fourth ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio;
and for each predictable value, taking the product of the actual flow related index and the third ratio corresponding to the predictable value and the product of the actual flow related index and the fourth ratio corresponding to the predictable value as end values of an interval, and obtaining an alternative interval corresponding to the moment after the release represented by the predictable value.
In a third aspect, an embodiment of the present invention provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when executed by a processor, the computer program implements the steps of the method for predicting a flow related indicator according to the first aspect.
Aiming at the prior art, the invention has the following advantages:
the candidate reference videos with the similarity meeting the first preset condition with the video to be estimated can be determined according to the relevant information of the video to be estimated and the relevant information of each candidate reference video, target videos are obtained, and estimated flow relevant indexes corresponding to different moments after the video to be estimated is released are determined based on flow relevant indexes corresponding to different moments after the video to be estimated is released. According to the embodiment of the invention, the flow related indexes corresponding to different moments after the video to be predicted is released can be predicted according to the video similar to the video to be predicted, so that the flow related indexes of the video to be predicted can be predicted more accurately. .
Furthermore, the alternative intervals of the flow related indexes corresponding to the videos to be estimated at different moments after the videos are issued can be determined based on the flow related indexes corresponding to the target videos at different moments after the videos are issued, so that a user can conveniently reselect one flow related index for the subsequent moment from the reasonable interval according to the actual condition when the actual condition does not accord with the condition represented by the estimated flow related indexes by determining the alternative intervals corresponding to the videos to be estimated at different moments after the videos are issued, and the accuracy of the estimated flow related indexes can be further improved by reasonably adjusting the flow related indexes to the estimated flow related indexes at the subsequent moments.
The foregoing description is only an overview of the technical solutions of the present invention, and the embodiments of the present invention are described below in order to make the technical means of the present invention more clearly understood and to make the above and other objects, features, and advantages of the present invention more clearly understandable.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
fig. 1 is a flowchart illustrating steps of a method for estimating a traffic related indicator according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating steps of another method for estimating a traffic related indicator according to an embodiment of the present invention;
fig. 3 is a block diagram of a device for estimating a flow related indicator according to an embodiment of the present invention;
fig. 4 is a block diagram of another device for estimating a flow related indicator according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
Fig. 1 is a flowchart illustrating steps of a method for estimating a traffic related indicator according to an embodiment of the present invention, as shown in fig. 1, the method may include:
In the embodiment of the present invention, the related information may be a parameter that affects traffic related indicators of videos, and different related information has different attractions to viewers, so that the number of viewers attracted to videos with different related information is different, and thus the traffic related indicators of the videos are different, for example, two videos with different directors may have different numbers of viewers, and thus the traffic related indicators of the videos are different, where the related information may include a single type of information or a plurality of different types of information, for example, the type of information corresponding to the related information may include a director, an investment amount, a production company, and the like, and correspondingly, the related information may include a specific director, a specific production amount, and a specific production company, for example, assuming that the director of the video to be estimated is director a, the director is director B, the investment amount is 100 ten thousand, and the production company is company C, the related information of the video to be estimated may be "director: director A and lead actor: the lead actor B and the investment amount: 100 ten thousand, production company: company C ".
Further, the second preset condition may be that a time length from the release time to the current time satisfies a preset time length threshold, where the release time may be a time point when the candidate reference video is on-line on the video platform, and the preset time length threshold may be set according to an actual request, for example, the preset time length threshold may be 60 days, or for example, the preset time length threshold may also be 100 days.
Further, the first preset condition may be that the similarity between the video to be estimated and the video to be estimated is greater than a preset similarity threshold, where the preset similarity threshold may be set according to an actual request, for example, the preset similarity threshold may be 80%, and for example, the preset similarity threshold may also be 90%, which is not limited in this embodiment of the present invention.
In the embodiment of the present invention, the different time after the release indicates a time point with the release time as a starting point and different time lengths from the release time, for example, a flow related index corresponding to the 1 st day, a flow related index corresponding to the 5 th day, a flow related index corresponding to the 10 th day, and the like of the video to be predicted may be determined in a time unit of day. Further, the target video is a candidate reference video which is selected based on the relevant information and is similar to the video to be estimated, and the relevant flow rate index of the video is in large relation with the relevant information of the video, so that the flow rate relevant index generated after the video to be estimated is released can be considered to be similar to the flow rate relevant index of the target video, and the flow rate relevant index of the video to be estimated can be estimated based on the target video.
In summary, the traffic related index estimation method provided in the embodiment of the present invention may determine, according to the related information of the video to be estimated and the related information of each candidate reference video, a candidate reference video whose similarity to the video to be estimated satisfies a first preset condition, to obtain a target video, where the candidate reference video is a video whose release time satisfies a second preset condition, and then, may determine, based on the traffic related indexes corresponding to different moments after release of each target video, the estimated traffic related indexes corresponding to different moments after release of the video to be estimated, where the different moments after release represent time nodes with different durations from the release moment. According to the embodiment of the invention, the flow related indexes corresponding to different moments after the video to be predicted is released can be predicted according to the video similar to the video to be predicted, so that the flow related indexes of the video to be predicted can be predicted more accurately. Further, because only fixed flow related indexes can be estimated for the video to be estimated in the prior art, a fixed operation strategy can be set for the video to be estimated only based on the estimated flow related indexes in the prior art, and in the embodiment of the invention, corresponding operation strategies can be set for different stages of the video after being released in advance by estimating the flow related indexes of the video to be estimated at different stages after being released, so that the popularization effect of the video to be estimated is improved.
Fig. 2 is a flowchart illustrating steps of another method for estimating a traffic related indicator according to an embodiment of the present invention, as shown in fig. 2, the method may include:
In this step, the relevant information of the video to be estimated and the relevant information of each candidate reference video may be obtained in advance before this step, so that by obtaining the relevant information of the video to be estimated and the relevant information of each candidate reference video in advance, the operation of determining the target video according to the relevant information can be directly executed according to the relevant information obtained in advance, the operation efficiency of determining the target video can be further improved, and the time consumed for realizing the estimation of the flow relevant index can be reduced. Of course, when the operation of determining the target video is performed, the related information of the video to be estimated and the related information of each candidate reference video may be obtained again, which is not limited in the embodiment of the present invention.
Specifically, the video to be estimated and each candidate reference video may be respectively used as a video to be processed, and the following operations indicated in steps a to C are performed:
and step A, acquiring information corresponding to the video to be processed from different data sources according to the type of the information contained in the related information, and obtaining multiple spare related information corresponding to the video to be processed.
In this step, the data source may be a data center of a network platform, for example, a background server of the network platform, and further, in an actual application scenario, it may occur that a plurality of network platform parties all have a copyright of a video, that is, the data sources corresponding to the plurality of network platform parties include related information of the same video, but problems that the related information may be inaccurate and the related information may be missing in the data source corresponding to the network platform party may occur. Specifically, the data source with the direct access right can be obtained by a direct access mode, and the data source without the direct access right can be obtained by a network crawling mode.
And B, carrying out duplicate removal processing on an information set formed by the multiple spare related information to obtain a target set.
In this step, each piece of backup related information may be combined to obtain an information set composed of the plurality of backup related information, for example, assuming that the related information includes director, player, playing platform, investment amount and 5 kinds of information of a production company, the director information, the player information, the investment amount information and the production company information corresponding to the video to be processed may be obtained from each data source, and accordingly, the information set may include the director information, the player information, the playing platform information, the investment amount information and the production company information obtained from each data source, specifically, each piece of backup related information may be respectively used as a candidate set, a union of the plurality of candidate sets is used as an information set, and then, the backup related information included in the information set may be subjected to deduplication processing, specifically, for the same type of information, only one of the information with the same content is retained, for example, for the information with the type of director, assuming that the information set includes director a, and director B, by the deduplication process, 3 directors a may be deleted, and directors a and directors B may be retained.
And step C, screening the information included in the target set according to the designated confidence degrees of different data sources, and taking the information included in the screened target set as the related information of the video to be processed.
Since there may be errors in data in different data sources, specific contents of information of the type existing in different data sources may be different for information of the same type, which may result in that a target set obtained after deduplication includes information of the same type and different represented contents, for example, for information of the type director, the target set may include a director a and a director B, in an embodiment of the present invention, a specified confidence may be set for each data source in advance according to the accuracy of the data sources in different data sources, and accordingly, for information of the same type and different represented contents, information with the largest specified confidence of the corresponding data source may be retained, and if the specification of the data source corresponding to the director a is greater than the specified confidence of the data source corresponding to the director B, the director B may be deleted, only director a is retained. Further, in the embodiment of the present invention, information included in the filtered target set may be stored in a structured manner, for example, the information is stored in a form of information of a type corresponding to the type, so as to facilitate reading of the information in a subsequent process. Further, in the embodiment of the present invention, the video to be estimated and the candidate reference video are both used as the video to be processed, so in this step, the related information of the video to be estimated and the related information of the candidate reference video can be obtained by obtaining the related information of each video to be processed.
In the embodiment of the invention, the alternative relevant information of the video to be processed is obtained from a plurality of data sources, the duplication of the alternative relevant information is removed, and the information with conflict is screened and corrected, so that the relevant information of the video to be processed is obtained, the obtained relevant information is more accurate and comprehensive, and the representativeness of the characteristics generated based on the relevant information in the subsequent steps is improved.
Further, step 201 may include the following steps 2011 to 2013:
and 2011, respectively generating the characteristics of the video to be estimated and the characteristics of each candidate reference video based on the relevant information of the video to be estimated and the relevant information of each candidate reference video.
In this step, the video to be estimated and each candidate reference video may be used as a video to be processed, and for each type of information in the related information of each video to be processed, if the number of selectable objects corresponding to the type of information is not greater than a preset threshold, the operation of generating the feature may be implemented through the following substeps (1) to (4):
substep (1): setting the value of an element corresponding to an optional object matched with the object represented by the information in the designated element sequence as a first mark value; the selectable objects corresponding to the type of the information correspond to the elements contained in the specified element sequence one to one.
Specifically, the selectable objects corresponding to the type of the information refer to objects representing different contents corresponding to the type, where the selectable objects corresponding to the type of the information may be predefined manually, and taking the type as a director for example, assuming that there are 100 directors in total, the selectable objects corresponding to the type of the information are the 100 directors, and the number of the selectable objects corresponding to the type of the information is 100, further, the preset threshold may be set according to an actual requirement, since the selectable objects corresponding to the type of the information correspond to the elements included in the specified element sequence one-to-one, therefore, to avoid that the specified element sequence includes more elements, the efficiency of generating the feature is improved, and the dimension of the generated feature is reduced, when the number of the corresponding selectable objects is small, the feature may be generated based on the specified element sequence, and accordingly, a smaller preset threshold may be set, for example, the preset threshold may be set to 4. Further, the first flag value may be set according to actual conditions, for example, the first flag value may be 1.
Substep (2): and setting the values of other elements in the specified element sequence as second mark values.
In this step, the second flag value may be a value different from the first flag value, which is preset, for example, the second flag value may be 0, and it is assumed that the type of the information is a subject of the video, and the subject corresponds to 4 kinds of selectable objects: the method includes the steps that the subjects a, the subjects b, the subjects c and the subjects d are defined in advance by developers, correspondingly, the specified element sequence corresponding to the type can include 4 elements of objects corresponding to the subjects a, the subjects b, the subjects c and the subjects d, if the subject information of the video to be processed is the subjects a and the subjects b, the optional objects matched with the objects represented by the information can be regarded as the subjects a and the subjects b, and therefore, the value of the element corresponding to the subject a and the value of the element corresponding to the subject b in the specified element sequence can be set to be 1, and the value of the element corresponding to the subject c and the value of the element corresponding to the subject d can be set to be 0.
Suppose the type of information is the television station playing the video, and the television station corresponds to 4 selectable objects: the method comprises the steps that a television station a, a television station b, a television station c and a television station d are respectively used as television station information of a video to be processed, and then the optional object matched with the object represented by the information can be considered as the television station a and the television station c, so that the value of an element corresponding to the television station a and the value of an element corresponding to the television station c in an appointed element sequence can be set to be 1, and the value of an element corresponding to the television station b and the value of an element corresponding to the television station d are set to be 0.
Suppose the type of information is a video website for playing the video, and the video website corresponds to 4 selectable objects: the video website information of the video to be processed is the video website b and the video website c, so that the value of the element corresponding to the video website b and the value of the element corresponding to the video website c in the specified element sequence can be set to 1, and the value of the element corresponding to the video website a and the value of the element corresponding to the video website d can be set to 0.
Assuming that the type of information is a play policy, the play policy corresponds to 4 kinds of selectable objects: the playing policy information of the video to be processed is the playing policy c and the playing policy d, so that the value of the element corresponding to the playing policy c and the value of the element corresponding to the playing policy d in the specified element sequence can be set to 1, and the value of the element corresponding to the playing policy a and the value of the element corresponding to the playing policy b are set to 0.
Assuming the type of information is investment operation cost, the investment operation is made to correspond to 4 optional objects: the cost range a, the cost range b, the cost range c, and the cost range d, where the investment operation cost information of the video to be processed falls within the cost range a, the value of the element corresponding to the cost range a in the specified element sequence may be set to 1, and the values of the elements corresponding to the cost range b, the cost range c, and the cost range d are all set to 0.
Substep (3): and taking the set designated element sequence as the characteristic corresponding to the information.
For example, taking the specified element sequence set for each type of information in the sub-step (2) as an example, assuming that the type of information is the subject of the video, the specified element sequence after setting may be: 1100 as a feature corresponding to the information; assuming that the type of information is the video station of the video play, the specified sequence of elements after the setting can be: 1010 as a feature corresponding to the information; assuming that the type of information is the video website on which the video is played, the specified sequence of elements after the setting can be: 0110 as the corresponding characteristic of the information; assuming that the type of information is a playback policy, the specified sequence of elements after setting can be: 0011 as a feature corresponding to the information; assuming the type of information is investment running cost, the specified sequence of elements after setting can be: 1000 as the characteristic corresponding to this information.
Substep (4): and combining the characteristics corresponding to each type of information to obtain the characteristics corresponding to the video to be processed.
In this step, a feature set formed by features corresponding to each type of information may be used as features corresponding to the video to be processed, that is, the features corresponding to the video to be processed include features corresponding to each type of information.
Further, if the number of selectable objects corresponding to the type of the information is greater than the preset threshold, the operation of generating the feature may be implemented by the following sub-step (1):
substep (1): and generating the characteristics of the video to be processed according to the network index of the object represented by the information.
In this step, the network index may be an average value, a median value, a maximum value, or a variance of the number of times that the object represented by the information is searched in the network, and accordingly, the network index of the object represented by the information may be determined according to the searched record of the object represented by the information in the network, and then the network index may be used as a feature of the information, and finally, features of all types of information included in the related information may be combined to be used as a feature of the video to be processed.
For example, assuming that the type of the information is a director of the video, a specific director indicated by the information, for example, "zhang san", and a network index for a preset time period may be used as a feature of the information. Assuming that the type of information is a production company of the video, a specific production company, for example, "dream company", represented by the information may be characterized by a network index for a preset time period. Assuming that the type of the information is the belonging series of the video, the network index of the specific belonging series represented by the information in a preset time period can be used as the characteristic of the information. Assuming that the type of the information is an adaptation of a video, a network index of a specific adaptation represented by the information for a preset time period may be used as a characteristic of the information. Of course, in practical applications, for information that itself represents a numerical value, the numerical value may be used as a feature of the information, and for example, for a video in which a plurality of sub videos exist, for example, a tv series, the number of sets of sub videos included in the video may be used as an "number of sets" feature of the information.
Step 2012, respectively calculating the euclidean distance between each candidate reference video and the video to be estimated according to the features of each candidate reference video and the features of the video to be estimated, and taking the euclidean distance as the similarity.
In this step, the euclidean distance between the candidate reference video and the video to be estimated is the similarity between the candidate reference video and the video to be estimated, and specifically, for each candidate reference video, the euclidean distance between the video to be estimated and the candidate reference video may be calculated by the following formula:
wherein D represents the euclidean distance between the candidate reference video and the video to be predicted, i represents the feature corresponding to the i-th type of information included in the related information, m represents the number of features, xXi represents the i-th feature of the video to be predicted, and yi represents the i-th feature of the candidate reference video.
And 2013, taking the candidate reference video with the corresponding Euclidean distance smaller than a preset distance threshold value as a target video.
In this step, the preset distance threshold may be set according to an actual situation, which is not limited in the embodiment of the present invention, and further, if the euclidean distance corresponding to the candidate reference video is smaller than the preset distance threshold, it may be considered that the candidate reference video is sufficiently similar to the video to be estimated, and the candidate reference video may be used as the target video.
In this step, the video to be estimated may be a video that has not been published, that is, the video to be estimated is a video that has been manufactured but has not been published to the network and cannot be selected by the user for viewing, and the video to be estimated may also be a video that has been published to the network, that is, the video to be estimated is a video that has been published to the network and has some traffic related indexes accumulated by the user for viewing, for example, the video a that has been published to the network 2 days ago and has the two accumulated traffic indexes n is a video that has been published.
Further, if the video to be estimated is a published video, step 202 may be implemented by the following step 2021:
step 2021, determining a predicted flow related index according to an actual flow related index corresponding to the video to be estimated and a flow related index corresponding to each target video at different moments after release.
Specifically, step 2021 may include steps 2021a to 2021 d:
step 2021a, for each target video, according to the flow related indexes corresponding to different moments after the target video is published, constructing a time and flow related index curve of the target video in a coordinate system formed by a first axis and a second axis.
In this step, the first axis may represent different times after the distribution, and for example, the first axis may be a scale of the first axis on the day basis, on the day 1, day 2, and day 3, day … x after the distribution, and the second axis may represent the flow rate related index, where the second axis may be a vertical axis when the first axis is a horizontal axis, and the second axis may be a horizontal axis when the first axis is a vertical axis. Specifically, taking the first axis as the horizontal axis and the second axis as the vertical axis as an example, a coordinate system may be constructed first, then points represented by the flow related index corresponding to each time after the target video is released are described in the coordinate system, and finally, the points are connected in sequence, so that a time and flow related index curve of the target video can be obtained.
Step 2021b, performing normalization processing on the time-flow related index curve according to the key time corresponding to the target video and the flow related index corresponding to the key time.
In this step, the time interval between the key time and the release time of the target video is equal to a key time, where the key time may be preset according to the actual situation of the target video, for example, when the target video is a series video including a plurality of sub-videos, the time consumed from releasing the first sub-video to releasing the last sub-video may be used as the key time, assuming that the target video is a television play including 30 episodes, the time consumed from releasing the first episode to releasing the last episode, that is, the broadcast time of the television play may be used as the key time, assuming that all episodes of the television play are released in 15 days, then the key time may be determined to be 15 days, further, when the target video is a single video, for example, the target video is a movie, that a professional may perform the following movie situation, the "hot-air period" of the movie is determined, and then the duration of the hot-air period is set to a critical duration, for example, the critical duration may be set to 20 days, and correspondingly, the critical moment may be the 20 th day after the release of the movie.
Further, when the normalization processing is performed: for each coordinate point in the time-flow related index curve, a value corresponding to a first axis in the coordinate point may be set as a quotient of the value and the critical duration, and a value corresponding to a second axis in the coordinate point may be set as a quotient of the value and the flow related index corresponding to the critical time. Therefore, the numerical value of the first axis is normalized through the key time length, the numerical value of the first axis is normalized through the flow related indexes corresponding to the key time, the influence caused by the difference of the hot broadcast periods or broadcast time lengths of different videos can be eliminated, the time corresponding to each target video and the flow related index curve are enabled to have the same standard, meanwhile, the data in the curve are enabled to be more concentrated, and the problem that the data are too sparse is avoided.
Step 2021c, generating an average curve of the time and flow related indexes according to the normalized time and flow related index curve of each target video.
In this step, the multiple normalized time and flow rate related index curves may be aggregated, specifically, an average value of coordinate points having the same value corresponding to the first axis in the multiple curves is calculated to obtain an average coordinate point, and then a curve formed by all the average coordinate points once is used as an average curve of the time and flow rate related index.
Step 2021d, calculating the predicted flow related index according to the actual flow related index corresponding to the video to be estimated, the published time length corresponding to the actual flow related index, the average curve of the time and flow related index, and the key time length.
Specifically, step 2021d may include the following substeps (1) to (4):
substep (1): and calculating the ratio of the published time length to the key time length corresponding to the video to be estimated to obtain a first ratio.
In this step, the published time length is a time interval between a time corresponding to the actual flow related index and a publishing time of the video to be predicted, and if the actual flow related index corresponding to the video to be predicted is the actual flow related index corresponding to the 2 nd day after the publishing of the video to be predicted, the published time length can be determined to be 2 days, and the key time length corresponding to the video to be predicted can be set for the video to be predicted in advance.
Substep (2): and searching the flow related indexes corresponding to the first ratio from the average curve of the time and flow related indexes, and searching the flow related indexes corresponding to each predictable value.
In this step, since the actual flow related index is the actual flow related index corresponding to the value not greater than the first ratio, the flow related index corresponding to the time indicated by the value greater than the first ratio can be estimated only, so as to avoid performing meaningless estimation and avoid wasting processing resources. Accordingly, the predictable value may be a value in the first axis greater than the first ratio.
Specifically, a coordinate point whose value corresponding to the first axis is the first ratio may be determined from the average curve, and then a value corresponding to the second axis included in the coordinate point is used as the flow related index corresponding to the first ratio.
Substep (3): and respectively calculating the ratio of the flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio to obtain a second ratio corresponding to each predictable value.
In this step, it is assumed that the flow related index corresponding to the predictable value is p and the flow related index corresponding to the first ratio is q, and then the second ratio can be represented as p/q.
Substep (4): and for a second ratio corresponding to each predictable value, calculating the product of the actual flow related index and the second ratio to obtain a predicted flow related index corresponding to the moment of the video to be predicted after the release represented by the predictable value.
In this step, the second ratio may represent a multiple relationship between the flow related index corresponding to the time after the release of the video to be estimated, which is represented by an estimable value, and the actual flow related index, for example, if the second ratio is 5, the released time is 2 days, and the time after the release of the video to be estimated, which is represented by the estimable value, is the 5 th day after the release of the video to be estimated, it may be considered that the flow related index corresponding to the 5 th day after the release of the video to be estimated may reach 5 times the flow related index of the 2 nd day after the release. Correspondingly, the product of the actual flow related index and the second ratio can be used as the predicted flow related index corresponding to the time length of the video to be predicted after the release of the predictable numerical value.
In the embodiment of the invention, when the video to be estimated is a published video with actual flow related indexes, the actual flow related indexes of the video to be estimated and historical flow related indexes corresponding to target videos similar to the video to be estimated at different moments after the publication are combined to estimate the flow related indexes corresponding to the video to be estimated at different moments after the publication.
Further, if the video to be estimated is an unpublished video, step 202 may be implemented by steps 2022 to 2023 as follows:
step 2022, calculating average flow related indexes corresponding to different moments after the release according to the flow related indexes corresponding to different moments after the release of each target video.
For example, the sum of the traffic related indexes corresponding to each target video on the 1 st day after the target video is published may be calculated in units of days, then the ratio of the sum to the number of the target videos is calculated to obtain the average traffic related index corresponding to the 1 st day after the target video is published, and so on, the sum of the traffic related indexes corresponding to each target video on the X th day after the target video is published may be calculated, then the ratio of the sum to the number of the target videos is calculated to obtain the average traffic related index corresponding to the X th day after the target video is published.
Step 2023, using the average traffic related indexes corresponding to different moments after the release as estimated traffic related indexes corresponding to different moments after the release of the video to be estimated.
For example, the average traffic related index corresponding to the 1 st day after the release may be used as the estimated traffic related index corresponding to the 1 st day after the release of the video to be estimated, and the average traffic related index corresponding to the X th day after the release may be used as the estimated traffic related index corresponding to the X th day after the release of the video to be estimated. It should be noted that, of course, when the video to be estimated is an unpublished video, a "relevant index of actual flow" may also be estimated by a professional, and the estimation is implemented by using the method in step 2021, which is not limited in the embodiment of the present invention.
Specifically, if the video to be estimated is a published video, this step may be implemented by the following step 2031:
step 2031, determining alternative intervals of the traffic related indexes corresponding to different moments after the release of the video to be estimated according to the actual traffic related indexes corresponding to the video to be estimated and the traffic related indexes corresponding to different moments after the release of each target video.
In the embodiment of the invention, because the actual flow related index can represent the actual attraction degree of the video to be estimated to the user, the actual flow related index generated in a period of time by combining the video to be estimated is estimated, the alternative interval of the estimated flow related index can be improved to a certain extent, and the estimation effect is further improved.
Specifically, step 2031 can be implemented by the following steps, sub-step (1) to sub-step (4):
substep (1): and searching a maximum flow related index and a minimum flow related index corresponding to the first ratio according to the normalized time and flow related index curve of each target video, and searching a maximum flow related index and a minimum flow related index corresponding to each predictable value.
In this step, for a normalized time-to-flow related index curve of each target video, a coordinate point whose value corresponding to the first axis is a first ratio may be determined from the curve, and then a value corresponding to the second axis included in the coordinate point is obtained to obtain a plurality of flow related indexes corresponding to the first ratio, and then a maximum value of the plurality of flow related indexes is used as a maximum flow related index corresponding to the first ratio, and a minimum value of the plurality of flow related indexes is used as a minimum flow related index corresponding to the first ratio.
Further, for each predictable value, a coordinate point corresponding to the first axis as the predictable value may be determined from each curve, and then a value corresponding to the second axis included in the coordinate point is obtained to obtain a plurality of flow related indicators corresponding to the predictable value, and then a maximum value of the plurality of flow related indicators is used as a maximum flow related indicator corresponding to the predictable value, and a minimum value of the plurality of flow related indicators is used as a minimum flow related indicator corresponding to the predictable value, so as to obtain a maximum flow related indicator and a minimum flow related indicator corresponding to each predictable value.
Substep (2): and respectively calculating a third ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio, and a fourth ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio.
In this step, if the maximum flow related index corresponding to the predictable value is r and the flow related index corresponding to the first ratio is q, the third ratio may be r/q, and further, if the minimum flow related index corresponding to the predictable value is s, the fourth ratio may be s/q.
Substep (3): and for each predictable value, taking the product of the actual flow related index and the third ratio corresponding to the predictable value and the product of the actual flow related index and the fourth ratio corresponding to the predictable value as end values of an interval, and obtaining an alternative interval corresponding to the time length after the release represented by the predictable value.
In this step, the third ratio may represent a maximum multiple relationship between the flow related index corresponding to the moment after the video to be estimated is released, which is represented by an estimable value, and the actual flow related index, for example, if the third ratio is 6, the released time is 2 days, and the moment after the video to be estimated, which is represented by the estimable value, is 7 days after the video to be estimated is released, it may be considered that the flow related index corresponding to the 7 th day after the video to be estimated is released will reach 6 times as much as the flow related index of the 2 nd day after the video to be estimated is released, and accordingly, a product of the actual flow related index and the third ratio may be used as an upper limit end value of an alternative interval corresponding to the moment after the video to be estimated is released, which is represented by the estimable value. Further, a fourth ratio may represent a minimum multiple relationship between a flow related index corresponding to a time after the video to be predicted is released, which is represented by a predictable value, and an actual flow related index, for example, assuming that the fourth ratio is 4, a released time is 2 days, and a time after the video to be predicted, which is represented by the predictable value, is 7 days after the video to be predicted is released, it may be considered that the flow related index corresponding to the 7 th day after the video to be predicted is released will reach 4 times as much as the flow related index at the 2 nd day after the video is released, and accordingly, a product of the actual flow related index and the fourth ratio may be used as a lower limit value of an alternative interval corresponding to the time after the video to be predicted is released, which is represented by the predictable value. Finally, the interval formed by the upper limit end value and the lower limit end value can be used as an alternative interval of the traffic related index corresponding to the 7 th day after the release of the video to be predicted.
Further, if the video to be estimated is an unreleased video, this step may be implemented by the following steps 2032 to 2033:
step 2032, according to the traffic related indexes corresponding to different moments after the release of each target video, obtaining the maximum traffic related indexes and the minimum traffic related indexes corresponding to different moments after the release.
For example, the traffic related indicator corresponding to the x-th day after the release of each target video may be obtained in units of days, so as to obtain a plurality of traffic related indicators corresponding to the x-th day after the release, and then, a maximum value of the plurality of traffic related indicators is used as a maximum traffic related indicator corresponding to the x-th day after the release, and a minimum value of the plurality of traffic related indicators is used as a minimum traffic related indicator corresponding to the x-th day after the release.
Step 2033, regarding the maximum flow rate related index and the minimum flow rate related index corresponding to different times after the release, taking the maximum flow rate related index and the minimum flow rate related index as end values of the interval corresponding to the time, and obtaining an alternative interval corresponding to the time.
For example, the maximum traffic related indicator corresponding to the x-th day may be used as the upper limit end value of the candidate interval, the minimum traffic related indicator corresponding to the x-th day may be used as the lower limit end value of the candidate interval, and finally, an interval formed by the upper limit end value and the lower limit end value may be used as the candidate interval of the traffic related indicator corresponding to the x-th day after the video to be predicted is released.
In the embodiment of the invention, the alternative intervals corresponding to different moments after the video to be estimated is released are determined, so that when the actual condition does not accord with the condition represented by the estimated flow related index, namely the estimated flow related index at the subsequent moment is determined to be inaccurate, one flow related index can be reselected for the subsequent moment from the reasonable interval according to the actual condition, the estimated flow related index is reasonably adjusted to be the estimated flow related index at the subsequent moment, the accuracy of the estimated flow related index can be further improved, meanwhile, the reselected flow related index can be used as an adjustment standard to adjust the operation strategy of the video to be estimated, and the matched operation strategy can be formulated for the video. For example, if it is assumed that the number of times of playing the to-be-estimated video on the 7 th day after the to-be-estimated video is estimated to reach 3000 ten thousand before the to-be-estimated video is not published, and the number of times of playing the to-be-estimated video on the 15 th day after the to-be-estimated video is estimated to reach 5 hundred million before the to-be-estimated video is published, accordingly, more push advertising advertisements can be set for the video to improve the popularity of the video, but the number of times of playing the to-be-estimated video on the 7 th day after the to-be-estimated video is published is only 500 ten thousand, and a large difference exists between the number of times of playing the to-be-estimated video and the distance of 3000 ten thousand, at this time, the number of times of playing the to-be-estimated video on the 15 th day after the to reach 5 hundred million, and further, a value smaller than 5 can be selected from a reasonable interval corresponding.
In summary, according to the traffic-related index estimation method provided by the embodiment of the present invention, the candidate reference videos whose similarity to the video to be estimated satisfies the first preset condition are determined according to the related information of the video to be estimated and the related information of each candidate reference video, so as to obtain the target video, then, the estimated traffic-related indexes corresponding to the video to be estimated at different moments after being released are determined based on the traffic-related indexes corresponding to the target videos at different moments after being released, and finally, the alternative intervals of the traffic-related indexes corresponding to the video to be estimated at different moments after being released are determined based on the traffic-related indexes corresponding to the target videos at different moments after being released. In the embodiment of the invention, the flow related indexes corresponding to different moments after the video to be predicted is published can be predicted according to the video similar to the video to be predicted, so that the flow related indexes of the video to be predicted can be predicted more accurately, corresponding operation strategies are set for different stages after the video is published in advance, the popularization effect of the video to be predicted is improved, meanwhile, when the actual condition does not accord with the condition represented by the predicted flow related indexes by the actual condition, a user can conveniently reselect a flow related index for the subsequent moment from the reasonable interval according to the actual condition by determining the alternative interval corresponding to the different moments after the video to be predicted is published, the accuracy of the predicted flow related indexes is further improved by reasonably adjusting the reselected flow related indexes to the predicted flow related indexes at the subsequent moment, and meanwhile, the reselected flow related indexes are used as an adjustment standard, and adjusting the operation strategy of the video to be estimated so as to ensure that a matched operation strategy can be formulated for the video.
Fig. 3 is a device for estimating a flow related indicator according to an embodiment of the present invention, as shown in fig. 3, the device 30 may include:
the first determining module 301 is configured to determine, according to the relevant information of the video to be estimated and the relevant information of each candidate reference video, a candidate reference video whose similarity to the video to be estimated meets a first preset condition, so as to obtain a target video; the candidate reference video is a video whose release time meets a second preset condition.
A second determining module 302, configured to determine estimated traffic related indexes corresponding to different moments after release of the to-be-estimated video based on traffic related indexes corresponding to different moments after release of each target video; and the different moments after the release represent time points which take the release moment as a starting point and are different from the release moment in duration.
In summary, in the traffic related indicator estimation apparatus provided in the embodiment of the present invention, the first determining module may determine, according to the related information of the video to be estimated and the related information of each candidate reference video, the candidate reference video whose similarity to the video to be estimated satisfies the first preset condition to obtain the target video, and then the second determining module may determine, based on the traffic related indicators corresponding to different moments after the release of each target video, the estimated traffic related indicators corresponding to different moments after the release of the video to be estimated. According to the embodiment of the invention, the flow related indexes corresponding to different moments after the video to be predicted is released can be predicted according to the video similar to the video to be predicted, so that the flow related indexes of the video to be predicted can be predicted more accurately, corresponding operation strategies can be formulated for different stages after the video is released in advance, and the popularization effect of the video to be predicted is improved.
Fig. 4 is a block diagram of another flow related indicator estimation apparatus according to an embodiment of the present invention, and as shown in fig. 4, the apparatus 40 may include:
the first determining module 401 is configured to determine, according to the relevant information of the video to be estimated and the relevant information of each candidate reference video, a candidate reference video whose similarity to the video to be estimated meets a first preset condition, so as to obtain a target video; the candidate reference video is a video whose release time meets a second preset condition.
A second determining module 402, configured to determine estimated traffic related indexes corresponding to different moments after release of the to-be-estimated video based on traffic related indexes corresponding to different moments after release of each target video; and the different moments after the release represent time points which take the release moment as a starting point and are different from the release moment in duration.
Optionally, the first determining module 401 includes:
the generation submodule is used for respectively generating the characteristics of the video to be estimated and the characteristics of each candidate reference video based on the relevant information of the video to be estimated and the relevant information of each candidate reference video;
the first calculation submodule is used for respectively calculating the Euclidean distance between each candidate reference video and the video to be estimated according to the characteristics of each candidate reference video and the characteristics of the video to be estimated, and the Euclidean distance is used as the similarity;
and the first determining submodule is used for taking the candidate reference video with the corresponding Euclidean distance smaller than a preset distance threshold value as the target video.
Optionally, the related information includes different types of information;
the device further comprises:
an obtaining module 403, configured to use the video to be estimated and each candidate reference video as a video to be processed, and obtain information corresponding to the video to be processed from different data sources according to types of information included in the related information, to obtain multiple pieces of standby related information corresponding to the video to be processed;
a duplicate removal module 404, configured to perform duplicate removal processing on an information set formed by the multiple pieces of standby related information to obtain a target set;
the screening module 405 is configured to screen information included in the target set according to the specified confidence degrees of different data sources, and use the information included in the screened target set as related information of the video to be processed.
Optionally, the generating sub-module is configured to:
for each type of information in the related information of each video to be processed, the following operations are respectively executed:
if the number of the selectable objects corresponding to the type of the information is not larger than a preset threshold value, setting the value of an element corresponding to the selectable object matched with the object represented by the information in the designated element sequence as a first mark value; setting values of other elements in the specified element sequence as second marker values; taking the set designated element sequence as a characteristic corresponding to the information; combining the characteristics corresponding to each type of information to obtain the characteristics corresponding to the video to be processed; the selectable objects corresponding to the type of the information correspond to the elements contained in the specified element sequence one by one; or,
and if the number of the selectable objects corresponding to the type of the information is larger than the preset threshold value, generating the characteristics of the video to be processed according to the network index of the object represented by the information.
Optionally, the second determining module 402 includes:
the second determining submodule is used for determining a predicted flow related index according to an actual flow related index corresponding to the video to be estimated and flow related indexes corresponding to different moments after the target video is released if the video to be estimated is a released video; or,
the second calculation submodule is used for calculating the average flow related indexes corresponding to different moments after the video is published according to the flow related indexes corresponding to different moments after the target video is published if the video to be estimated is a video which is not published; and the third determining submodule is used for taking the average flow related indexes corresponding to different moments after the release as the estimated flow related indexes corresponding to different moments after the release of the video to be estimated.
Optionally, the second determining sub-module includes:
the construction unit is used for constructing a time and flow related index curve of each target video in a coordinate system consisting of a first axis and a second axis according to flow related indexes corresponding to different moments after the target video is published; the first axis represents different moments after release, and the second axis represents flow related indicators;
the normalization unit is used for performing normalization processing on the time and flow related index curve according to a key moment corresponding to the target video and a flow related index corresponding to the key moment; the time interval between the key moment and the release moment of the target video is equal to the key duration;
the acquisition unit is used for acquiring an average curve of the time and flow related indexes according to the normalized time and flow related index curve of each target video;
and the calculating unit is used for calculating the predicted flow related index according to the actual flow related index corresponding to the video to be estimated, the published time length corresponding to the actual flow related index, the average curve of the time and flow related indexes and the key time length.
Optionally, the normalization unit is configured to:
and for each coordinate point in the curve of the time-flow related index, setting a numerical value corresponding to a first axis in the coordinate points as a quotient of the numerical value and the key duration, and setting a numerical value corresponding to a second axis in the coordinate points as a quotient of the numerical value and the flow related index corresponding to the key moment.
Optionally, the published duration is a time interval between a moment corresponding to the actual flow related index and the publishing moment of the video to be estimated;
the computing unit is configured to:
calculating the ratio of the published time length to the key time length corresponding to the video to be estimated to obtain a first ratio;
searching the flow related indexes corresponding to the first ratio from the average curve of the time and flow related indexes, and searching the flow related indexes corresponding to each predictable value; the predictable value represents a value in the first axis that is greater than the first ratio;
respectively calculating the ratio of the flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio to obtain a second ratio corresponding to each predictable value;
and for a second ratio corresponding to each predictable value, calculating the product of the actual flow related index and the second ratio to obtain a predicted flow related index corresponding to the moment of the video to be predicted after the release represented by the predictable value.
Optionally, the apparatus 40 further includes:
and the third determining module is used for determining alternative intervals of the flow related indexes corresponding to different moments of the video to be estimated after the video is published based on the flow related indexes corresponding to different moments of each target video after the video is published.
Optionally, the third determining module includes:
a fourth determining submodule, configured to determine, if the video to be predicted is a published video, alternative intervals of the traffic related indexes corresponding to different moments after the release of the video to be predicted according to an actual traffic related index corresponding to the video to be predicted and traffic related indexes corresponding to different moments after the release of each target video;
the obtaining submodule is used for obtaining maximum flow related indexes and minimum flow related indexes corresponding to different moments after the target videos are published according to the flow related indexes corresponding to the different moments after the target videos are published if the videos to be estimated are unpublished videos; and a fifth determining submodule, configured to, for a maximum traffic related indicator and a minimum traffic related indicator corresponding to different times after the release, obtain an alternative interval corresponding to the time by using the maximum traffic related indicator and the minimum traffic related indicator as end values of an interval corresponding to the time.
Optionally, the fourth determining sub-module is configured to:
searching a maximum flow related index and a minimum flow related index corresponding to the first ratio according to the normalized time and flow related index curve of each target video, and searching a maximum flow related index and a minimum flow related index corresponding to each predictable value;
respectively calculating a third ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio, and a fourth ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio;
and for each predictable value, taking the product of the actual flow related index and the third ratio corresponding to the predictable value and the product of the actual flow related index and the fourth ratio corresponding to the predictable value as end values of an interval, and obtaining an alternative interval corresponding to the moment after the release represented by the predictable value.
In summary, in the traffic related index estimation apparatus provided in the embodiment of the present invention, the first determining module may determine, according to the related information of the video to be estimated and the related information of each candidate reference video, the candidate reference video whose similarity to the video to be estimated satisfies the first preset condition, to obtain the target video, then, the second determining module may determine, based on the traffic related indexes corresponding to different moments after the release of each target video, the estimated traffic related indexes corresponding to different moments after the release of the video to be estimated, and finally, the third determining module may determine, based on the traffic related indexes corresponding to different moments after the release of each target video, the candidate intervals of the traffic related indexes corresponding to different moments after the release of the video to be estimated. In the embodiment of the invention, the flow related indexes corresponding to different moments after the video to be predicted is published can be predicted according to the video similar to the video to be predicted, so that the flow related indexes of the video to be predicted can be predicted more accurately, corresponding operation strategies are set for different stages after the video is published in advance, the popularization effect of the video to be predicted is improved, meanwhile, when the actual condition does not accord with the condition represented by the predicted flow related indexes by the actual condition, a user can conveniently reselect a flow related index for the subsequent moment from the reasonable interval according to the actual condition by determining the alternative interval corresponding to the different moments after the video to be predicted is published, the estimated flow related indexes are reasonably adjusted to be the estimated flow related indexes at the subsequent moment, the accuracy of the estimated flow related indexes can be further improved, and meanwhile, the reselected flow related indexes can be used as an adjustment standard, and adjusting the operation strategy of the video to be estimated so as to ensure that a matched operation strategy can be formulated for the video.
For the above device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for the relevant points, refer to the partial description of the method embodiment.
Preferably, an embodiment of the present invention further provides a terminal, including a processor, a memory, and a computer program stored in the memory and capable of running on the processor, where the computer program, when executed by the processor, implements each process of the embodiment of the traffic related indicator estimation method, and can achieve the same technical effect, and details are not repeated here to avoid repetition.
The embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the computer program implements the processes of the above embodiment of estimating the flow related index, and can achieve the same technical effect, and in order to avoid repetition, the detailed description is omitted here. The computer-readable storage medium may be a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk.
The embodiments in the present specification are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
As is readily imaginable to the person skilled in the art: any combination of the above embodiments is possible, and thus any combination between the above embodiments is an embodiment of the present invention, but the present disclosure is not necessarily detailed herein for reasons of space.
The flow related indicator prediction methods provided herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general purpose systems may also be used with the teachings herein. The structure required to construct a system incorporating aspects of the present invention will be apparent from the description above. Moreover, the present invention is not directed to any particular programming language. It is appreciated that a variety of programming languages may be used to implement the teachings of the present invention as described herein, and any descriptions of specific languages are provided above to disclose the best mode of the invention.
In the description provided herein, numerous specific details are set forth. It is understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.
Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the invention and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be interpreted as reflecting an intention that: that the invention as claimed requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.
Those skilled in the art will appreciate that the modules in the device in an embodiment may be adaptively changed and disposed in one or more devices different from the embodiment. The modules or units or components of the embodiments may be combined into one module or unit or component, and furthermore they may be divided into a plurality of sub-modules or sub-units or sub-components. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or elements of any method or apparatus so disclosed, may be combined in any combination, except combinations where at least some of such features and/or processes or elements are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.
Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features included in other embodiments, rather than other features, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the claims, any of the claimed embodiments may be used in any combination.
The various component embodiments of the invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that a microprocessor or Digital Signal Processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components of the flow related indicator estimation method according to embodiments of the present invention. The present invention may also be embodied as apparatus or device programs (e.g., computer programs and computer program products) for performing a portion or all of the methods described herein. Such programs implementing the present invention may be stored on computer-readable media or may be in the form of one or more signals. Such a signal may be downloaded from an internet website or provided on a carrier signal or in any other form.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The usage of the words first, second and third, etcetera do not indicate any ordering. These words may be interpreted as names.
Claims (21)
1. A method for predicting a flow related index is characterized by comprising the following steps:
determining candidate reference videos with the similarity to the video to be estimated meeting a first preset condition according to the relevant information of the video to be estimated and the relevant information of each candidate reference video to obtain a target video; the candidate reference video is a video meeting a second preset condition at the release moment;
determining estimated flow related indexes corresponding to different moments of the video to be estimated after release based on the flow related indexes corresponding to the different moments of each target video after release; different moments after the release represent time points which take the release moment as a starting point and have different time lengths from the release moment;
wherein the method further comprises:
determining alternative intervals of the flow related indexes corresponding to different moments of the video to be estimated after the video is published based on the flow related indexes corresponding to different moments of each target video after the video is published;
and under the condition that the actual flow related index after the video to be estimated is released does not accord with the estimated flow related index, re-determining the estimated flow related index according to the alternative interval.
2. The method according to claim 1, wherein the determining, according to the related information of the video to be predicted and the related information of each candidate reference video, the candidate reference video whose similarity with the video to be predicted satisfies a first preset condition to obtain the target video comprises:
respectively generating the characteristics of the video to be estimated and the characteristics of each candidate reference video based on the relevant information of the video to be estimated and the relevant information of each candidate reference video;
respectively calculating the Euclidean distance between each candidate reference video and the video to be estimated according to the characteristics of each candidate reference video and the characteristics of the video to be estimated to serve as the similarity;
and taking the candidate reference video with the corresponding Euclidean distance smaller than a preset distance threshold value as the target video.
3. The method of claim 2, wherein the related information comprises different types of information;
before determining the candidate reference video with the similarity to the video to be estimated meeting the first preset condition according to the relevant information of the video to be estimated and the relevant information of each candidate reference video, the method further comprises the following steps:
respectively taking the video to be estimated and each candidate reference video as a video to be processed, and acquiring information corresponding to the video to be processed from different data sources according to the type of the information contained in the related information to obtain multiple spare related information corresponding to the video to be processed;
carrying out duplicate removal processing on an information set formed by the multiple spare related information to obtain a target set;
and screening the information included in the target set according to the specified confidence degrees of different data sources, and taking the information included in the screened target set as the related information of the video to be processed.
4. The method according to claim 2, wherein the generating the feature of the candidate reference video and the feature of the candidate reference video based on the information about the to-be-predicted video and the information about the candidate reference video respectively comprises:
for each type of information in the related information of each video to be processed, the following operations are respectively executed:
if the number of the selectable objects corresponding to the type of the information is not larger than a preset threshold value, setting the value of an element corresponding to the selectable object matched with the object represented by the information in the designated element sequence as a first mark value; setting values of other elements in the specified element sequence as second marker values; taking the set designated element sequence as a characteristic corresponding to the information; combining the characteristics corresponding to each type of information to obtain the characteristics corresponding to the video to be processed; the selectable objects corresponding to the type of the information correspond to the elements contained in the specified element sequence one by one; or,
and if the number of the selectable objects corresponding to the type of the information is larger than the preset threshold value, generating the characteristics of the video to be processed according to the network index of the object represented by the information.
5. The method according to claim 1, wherein the determining the estimated traffic related indicators corresponding to the to-be-estimated video at different times after the release based on the traffic related indicators corresponding to each target video at different times after the release comprises:
if the video to be estimated is a published video, determining a predicted flow related index according to an actual flow related index corresponding to the video to be estimated and flow related indexes corresponding to different moments of each target video after the video to be estimated is published; or,
if the video to be estimated is a video which is not released, calculating average flow related indexes corresponding to different moments after releasing according to the flow related indexes corresponding to different moments after releasing of each target video; and taking the average flow related indexes corresponding to different moments after the release as estimated flow related indexes corresponding to different moments after the release of the video to be estimated.
6. The method according to claim 5, wherein the determining a predicted traffic related indicator according to an actual traffic related indicator corresponding to the video to be predicted and a traffic related indicator corresponding to each target video at different time after being released comprises:
for each target video, according to the corresponding flow related indexes of the target video at different moments after the target video is released, constructing a time and flow related index curve of the target video in a coordinate system formed by a first axis and a second axis; the first axis represents different moments after release, and the second axis represents flow related indicators;
according to the key moment corresponding to the target video and the flow related index corresponding to the key moment, carrying out normalization processing on the time and flow related index curve; the time interval between the key moment and the release moment of the target video is equal to the key duration;
acquiring an average curve of the time and flow related indexes according to the normalized time and flow related index curve of each target video;
and calculating the predicted flow related index according to the actual flow related index corresponding to the video to be estimated, the published time length corresponding to the actual flow related index, the average curve of the time and flow related indexes and the key time length.
7. The method according to claim 6, wherein the normalizing the time-flow related indicator curve according to a key moment corresponding to the target video and a flow related indicator corresponding to the key moment comprises:
and for each coordinate point in the curve of the time-flow related index, setting a numerical value corresponding to a first axis in the coordinate points as a quotient of the numerical value and the key duration, and setting a numerical value corresponding to a second axis in the coordinate points as a quotient of the numerical value and the flow related index corresponding to the key moment.
8. The method according to claim 7, wherein the published duration is a time interval from a time corresponding to the actual flow related indicator to a publishing time of the video to be predicted;
the calculating the predicted flow related index according to the actual flow related index corresponding to the video to be estimated, the published time length corresponding to the actual flow related index, the average curve of the time and flow related indexes and the key time length comprises:
calculating the ratio of the published time length to the key time length corresponding to the video to be estimated to obtain a first ratio;
searching the flow related indexes corresponding to the first ratio from the average curve of the time and flow related indexes, and searching the flow related indexes corresponding to each predictable value; the predictable value represents a value in the first axis that is greater than the first ratio;
respectively calculating the ratio of the flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio to obtain a second ratio corresponding to each predictable value;
and for a second ratio corresponding to each predictable value, calculating the product of the actual flow related index and the second ratio to obtain a predicted flow related index corresponding to the moment of the video to be predicted after the release represented by the predictable value.
9. The method according to claim 8, wherein the determining, based on the traffic related indicator corresponding to each target video at different time after being released, the alternative intervals of the traffic related indicator corresponding to the video to be predicted at different time after being released comprises:
if the video to be estimated is a published video, determining alternative intervals of the flow related indexes corresponding to different moments after the video to be estimated is published according to the actual flow related indexes corresponding to the video to be estimated and the flow related indexes corresponding to different moments after the video to be estimated is published;
if the video to be estimated is a video which is not released, acquiring maximum flow related indexes and minimum flow related indexes corresponding to different moments after releasing according to flow related indexes corresponding to different moments after releasing of each target video; and regarding the maximum flow rate related index and the minimum flow rate related index corresponding to different moments after the release, taking the maximum flow rate related index and the minimum flow rate related index as end values of a section corresponding to the moments, and obtaining a candidate section corresponding to the moments.
10. The method according to claim 9, wherein the determining, according to the actual traffic related indicator corresponding to the video to be predicted and the traffic related indicator corresponding to each target video at different time after being released, the alternative intervals of the traffic related indicators corresponding to the video to be predicted at different time after being released comprises:
searching a maximum flow related index and a minimum flow related index corresponding to the first ratio according to the normalized time and flow related index curve of each target video, and searching a maximum flow related index and a minimum flow related index corresponding to each predictable value;
respectively calculating a third ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio, and a fourth ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio;
and for each predictable value, taking the product of the actual flow related index and the third ratio corresponding to the predictable value and the product of the actual flow related index and the fourth ratio corresponding to the predictable value as end values of an interval, and obtaining an alternative interval corresponding to the moment after the release represented by the predictable value.
11. An apparatus for predicting a flow related indicator, the apparatus comprising:
the first determination module is used for determining candidate reference videos of which the similarity with the videos to be estimated meets a first preset condition according to the relevant information of the videos to be estimated and the relevant information of each candidate reference video to obtain target videos; the candidate reference video is a video meeting a second preset condition at the release moment;
the second determination module is used for determining estimated flow related indexes corresponding to different moments of the video to be estimated after the video is published based on the flow related indexes corresponding to different moments of each target video after the video is published; different moments after the release represent time points which take the release moment as a starting point and have different time lengths from the release moment;
wherein the apparatus further comprises:
the third determining module is used for determining alternative intervals of the flow related indexes corresponding to the videos to be estimated at different moments after the videos are published based on the flow related indexes corresponding to the target videos at different moments after the videos are published;
the apparatus is further configured to:
and under the condition that the actual flow related index after the video to be estimated is released does not accord with the estimated flow related index, re-determining the estimated flow related index according to the alternative interval.
12. The apparatus of claim 11, wherein the first determining module comprises:
the generation submodule is used for respectively generating the characteristics of the video to be estimated and the characteristics of each candidate reference video based on the relevant information of the video to be estimated and the relevant information of each candidate reference video;
the first calculation submodule is used for respectively calculating the Euclidean distance between each candidate reference video and the video to be estimated according to the characteristics of each candidate reference video and the characteristics of the video to be estimated, and the Euclidean distance is used as the similarity;
and the first determining submodule is used for taking the candidate reference video with the corresponding Euclidean distance smaller than a preset distance threshold value as the target video.
13. The apparatus of claim 12, wherein the related information comprises different types of information;
the device further comprises:
the acquisition module is used for respectively taking the video to be estimated and each candidate reference video as a video to be processed, and acquiring information corresponding to the video to be processed from different data sources according to the type of the information contained in the related information to obtain multiple spare related information corresponding to the video to be processed;
the duplication removing module is used for carrying out duplication removing processing on an information set formed by the multiple spare related information to obtain a target set;
and the screening module is used for screening the information included in the target set according to the specified confidence degrees of different data sources, and taking the information included in the screened target set as the related information of the video to be processed.
14. The apparatus of claim 12, wherein the generation submodule is configured to:
for each type of information in the related information of each video to be processed, the following operations are respectively executed:
if the number of the selectable objects corresponding to the type of the information is not larger than a preset threshold value, setting the value of an element corresponding to the selectable object matched with the object represented by the information in the designated element sequence as a first mark value; setting values of other elements in the specified element sequence as second marker values; taking the set designated element sequence as a characteristic corresponding to the information; combining the characteristics corresponding to each type of information to obtain the characteristics corresponding to the video to be processed; the selectable objects corresponding to the type of the information correspond to the elements contained in the specified element sequence one by one; or,
and if the number of the selectable objects corresponding to the type of the information is larger than the preset threshold value, generating the characteristics of the video to be processed according to the network index of the object represented by the information.
15. The apparatus of claim 11, wherein the second determining module comprises:
the second determining submodule is used for determining a predicted flow related index according to an actual flow related index corresponding to the video to be estimated and flow related indexes corresponding to different moments after the target video is released if the video to be estimated is a released video; or,
the second calculation submodule is used for calculating the average flow related indexes corresponding to different moments after the video is published according to the flow related indexes corresponding to different moments after the target video is published if the video to be estimated is a video which is not published; and the third determining submodule is used for taking the average flow related indexes corresponding to different moments after the release as the estimated flow related indexes corresponding to different moments after the release of the video to be estimated.
16. The apparatus of claim 15, wherein the second determining submodule comprises:
the construction unit is used for constructing a time and flow related index curve of each target video in a coordinate system consisting of a first axis and a second axis according to flow related indexes corresponding to different moments after the target video is published; the first axis represents different moments after release, and the second axis represents flow related indicators;
the normalization unit is used for performing normalization processing on the time and flow related index curve according to a key moment corresponding to the target video and a flow related index corresponding to the key moment; the time interval between the key moment and the release moment of the target video is equal to the key duration;
the acquisition unit is used for acquiring an average curve of the time and flow related indexes according to the normalized time and flow related index curve of each target video;
and the calculating unit is used for calculating the predicted flow related index according to the actual flow related index corresponding to the video to be estimated, the published time length corresponding to the actual flow related index, the average curve of the time and flow related indexes and the key time length.
17. The apparatus of claim 16, wherein the normalization unit is configured to:
and for each coordinate point in the curve of the time-flow related index, setting a numerical value corresponding to a first axis in the coordinate points as a quotient of the numerical value and the key duration, and setting a numerical value corresponding to a second axis in the coordinate points as a quotient of the numerical value and the flow related index corresponding to the key moment.
18. The apparatus according to claim 17, wherein the published duration is a time interval between a time corresponding to the actual flow related indicator and a publishing time of the video to be predicted;
the computing unit is configured to:
calculating the ratio of the published time length to the key time length corresponding to the video to be estimated to obtain a first ratio;
searching the flow related indexes corresponding to the first ratio from the average curve of the time and flow related indexes, and searching the flow related indexes corresponding to each predictable value; the predictable value represents a value in the first axis that is greater than the first ratio;
respectively calculating the ratio of the flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio to obtain a second ratio corresponding to each predictable value;
and for a second ratio corresponding to each predictable value, calculating the product of the actual flow related index and the second ratio to obtain a predicted flow related index corresponding to the moment of the video to be predicted after the release represented by the predictable value.
19. The apparatus of claim 18, wherein the third determining module comprises:
a fourth determining submodule, configured to determine, if the video to be predicted is a published video, alternative intervals of the traffic related indexes corresponding to different moments after the release of the video to be predicted according to an actual traffic related index corresponding to the video to be predicted and traffic related indexes corresponding to different moments after the release of each target video;
the obtaining submodule is used for obtaining maximum flow related indexes and minimum flow related indexes corresponding to different moments after the target videos are published according to the flow related indexes corresponding to the different moments after the target videos are published if the videos to be estimated are unpublished videos; and a fifth determining submodule, configured to, for a maximum traffic related indicator and a minimum traffic related indicator corresponding to different times after the release, obtain an alternative interval corresponding to the time by using the maximum traffic related indicator and the minimum traffic related indicator as end values of an interval corresponding to the time.
20. The apparatus of claim 19, wherein the fourth determination submodule is configured to:
searching a maximum flow related index and a minimum flow related index corresponding to the first ratio according to the normalized time and flow related index curve of each target video, and searching a maximum flow related index and a minimum flow related index corresponding to each predictable value;
respectively calculating a third ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio, and a fourth ratio of the maximum flow related index corresponding to each predictable value to the flow related index corresponding to the first ratio;
and for each predictable value, taking the product of the actual flow related index and the third ratio corresponding to the predictable value and the product of the actual flow related index and the fourth ratio corresponding to the predictable value as end values of an interval, and obtaining an alternative interval corresponding to the moment after the release represented by the predictable value.
21. A computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when executed by a processor, the computer program implements the method for estimating a flow related indicator according to any one of claims 1 to 10.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910082401.9A CN109963174B (en) | 2019-01-28 | 2019-01-28 | Flow related index estimation method and device and computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910082401.9A CN109963174B (en) | 2019-01-28 | 2019-01-28 | Flow related index estimation method and device and computer readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109963174A CN109963174A (en) | 2019-07-02 |
CN109963174B true CN109963174B (en) | 2021-06-22 |
Family
ID=67023502
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910082401.9A Active CN109963174B (en) | 2019-01-28 | 2019-01-28 | Flow related index estimation method and device and computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109963174B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112258214A (en) * | 2020-09-22 | 2021-01-22 | 北京达佳互联信息技术有限公司 | Video delivery method and device and server |
CN113347464B (en) * | 2021-04-30 | 2022-09-30 | 北京奇艺世纪科技有限公司 | Video playing flow prediction method and device, electronic equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103914475A (en) * | 2013-01-05 | 2014-07-09 | 腾讯科技(北京)有限公司 | Method, system and device for predicting video views |
CN107707964A (en) * | 2016-08-08 | 2018-02-16 | 华为软件技术有限公司 | The method and apparatus for predicting video content temperature |
CN108898415A (en) * | 2018-05-29 | 2018-11-27 | 北京奇艺世纪科技有限公司 | A kind of the flow index of correlation prediction technique and device of video collection of drama |
CN108989889A (en) * | 2018-08-09 | 2018-12-11 | 北京奇艺世纪科技有限公司 | A kind of video playing amount prediction technique, device and electronic equipment |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6766525B1 (en) * | 2000-02-08 | 2004-07-20 | Koninklijke Philips Electronics N.V. | Method and apparatus for evaluating television program recommenders |
US9912424B2 (en) * | 2016-05-27 | 2018-03-06 | The Nielsen Company (Us), Llc | Methods and apparatus to estimate ratings for media assets using social media |
-
2019
- 2019-01-28 CN CN201910082401.9A patent/CN109963174B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103914475A (en) * | 2013-01-05 | 2014-07-09 | 腾讯科技(北京)有限公司 | Method, system and device for predicting video views |
CN107707964A (en) * | 2016-08-08 | 2018-02-16 | 华为软件技术有限公司 | The method and apparatus for predicting video content temperature |
CN108898415A (en) * | 2018-05-29 | 2018-11-27 | 北京奇艺世纪科技有限公司 | A kind of the flow index of correlation prediction technique and device of video collection of drama |
CN108989889A (en) * | 2018-08-09 | 2018-12-11 | 北京奇艺世纪科技有限公司 | A kind of video playing amount prediction technique, device and electronic equipment |
Also Published As
Publication number | Publication date |
---|---|
CN109963174A (en) | 2019-07-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108062409B (en) | Live video abstract generation method and device and electronic equipment | |
US10133818B2 (en) | Estimating social interest in time-based media | |
US8995823B2 (en) | Method and system for content relevance score determination | |
US10623823B2 (en) | Method, system, and website server for playing website video | |
CN108683924B (en) | Video processing method and device | |
CN110941738B (en) | Recommendation method and device, electronic equipment and computer-readable storage medium | |
CN106454431B (en) | TV programme suggesting method and system | |
CN104768082A (en) | Audio/video playing information processing method and server | |
CN105786857B (en) | Method and system for improving video aggregation efficiency | |
CN112507163B (en) | Duration prediction model training method, recommendation method, device, equipment and medium | |
CN109963174B (en) | Flow related index estimation method and device and computer readable storage medium | |
CN104967690B (en) | A kind of information-pushing method and device | |
KR101541495B1 (en) | Apparatus, method and computer readable recording medium for analyzing a video using the image captured from the video | |
CN114782092A (en) | Advertisement putting method, device, equipment, storage medium and computer program | |
CN103442270A (en) | Method and device for acquiring viewing data of user | |
CN109688217B (en) | Message pushing method and device and electronic equipment | |
CN104581226A (en) | Method and device for recommending programs by aiming at current live stream | |
CN108764021B (en) | Cheating video identification method and device | |
JP2009200918A (en) | Program recording and playback apparatus | |
US8306992B2 (en) | System for determining content topicality, and method and program thereof | |
CN112804566A (en) | Program recommendation method, device and computer readable storage medium | |
CN109561350B (en) | User interest degree evaluation method and system | |
CN114996509A (en) | Method and device for training video feature extraction model and video recommendation | |
CN110139160B (en) | Prediction system and method | |
CN113938712B (en) | Video playing method and device and electronic equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |