CN104731950A - Video recommendation method and device - Google Patents
Video recommendation method and device Download PDFInfo
- Publication number
- CN104731950A CN104731950A CN201510150095.XA CN201510150095A CN104731950A CN 104731950 A CN104731950 A CN 104731950A CN 201510150095 A CN201510150095 A CN 201510150095A CN 104731950 A CN104731950 A CN 104731950A
- Authority
- CN
- China
- Prior art keywords
- video
- interest value
- user
- features
- video features
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/7867—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, title and artist information, manually generated time, location and usage information, user ratings
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Library & Information Science (AREA)
- Multimedia (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention discloses a video recommendation method and device. An operation behavior conducted on a video by a user and video identification of the operation behavior can be acquired, multiple video features corresponding to the acquired video identification are extracted locally, and all the extracted video features are processed in the following mode that according to preset feature weights corresponding to the video features and a preset behavior weight corresponding to the acquired operation behavior, interest values of the acquired video features are calculated, the interest values of the acquired video features and user information are correspondingly stored, and according to the stored interest values of the video features corresponding to the user information, recommended videos are determined and recommended to the user. When the videos are recommended to the user, the prior operation behavior of the user and the operation video features are considered, the recommended videos better arouse the interest of the user, and the video recommendation effect is improved.
Description
Technical field
The present invention relates to network application field, particularly relate to a kind of video recommendation method and device.
Background technology
Day by day universal along with computer network, people carry out the recreations such as viewing film, TV play on the net more and more.Video website often recommends video to user, stops temporally to use the head of a household in video website.
Prior art often according to current that watching or the just viewed video of user, recommends similar video to it." similar " mentioned here video, its classifying mode is relatively more rough, such as, if find that user have viewed an acrobatic fighting film, just recommends another acrobatic fighting film to it; If user have viewed the slice, thin piece that certain movie star acts the leading role, just to another slice, thin piece that it recommends this movie star to act the leading role.And sometimes user may open a video and watches, but closes it because of loseing interest in again immediately, but existing mode still can recommend similar video to user.Because there are these shortcomings above-mentioned, the video that the existing way of recommendation is recommended often does not meet user interest, and recommendation effect is poor.
Summary of the invention
The object of the embodiment of the present invention is to provide a kind of video recommendation method and device, to recommend video to user better.
For achieving the above object, the embodiment of the invention discloses a kind of video recommendation method, technical scheme is as follows:
A kind of video recommendation method, be applied to server, store the mark of video and the video features corresponding with the mark of video in described server this locality, described method comprises:
Obtain user for the operation behavior of video and described operation behavior for the mark of video;
The multiple video features corresponding with the mark of obtained video are extracted from this locality;
Each extracted video features is all handled as follows: according to the behavior weight corresponding with this video features characteristic of correspondence weight and the default described operation behavior with obtaining preset, calculate the interest value obtaining this video features;
By the interest value of obtained each video features and user profile corresponding stored;
The video that will recommend is determined and by the determined video recommendations that will recommend to user according to the interest value of stored each video features corresponding with user profile.
Preferably, described basis preset with this video features characteristic of correspondence weight and the default behavior weight corresponding with obtained operation behavior, calculate and obtain the interest value of this video features, comprising:
Determine the default feature weight corresponding with this video features and the default behavior weight corresponding with described operation behavior, wherein, all default with one the behavior weight of each operation behavior is corresponding; When obtained user is one for the operation behavior of video, by determined default feature weight and determined default behavior multiplied by weight, obtain the interest value of this video features; When obtained user is multiple for the operation behavior of video, by determined default feature weight respectively with determined each default behavior multiplied by weight, by each product addition, obtain the interest value of this video features.
Preferably, the described interest value according to the stored each video features corresponding with user profile is determined the video that will recommend and by the determined video recommendations that will recommend to before the step of user, is also comprised: by the rise time of the interest value of obtained each video features and user profile corresponding stored;
The described interest value according to the stored each video features corresponding with user profile is determined the video that will recommend and by the determined video recommendations that will recommend to user, being comprised:
According to the rise time of each interest value, process is weighted to each interest value;
The video that will recommend is determined and by the determined video recommendations that will recommend to user according to the interest value after weighting process.
Preferably, described to determine recommending with the interest value of each video features of user's corresponding stored according to stored video and by the determined video recommendations that will recommend to user, comprising:
According to the height of the interest value of each video features, select the high interest value video features of the first number;
For the high interest value video features of the first selected number, select the video of the second number from video library, wherein, each video in the video of described second number has at least one high interest value video features;
The video of described second number is recommended to user.
Preferably, described to determine recommending with the interest value of each video features of user's corresponding stored according to stored video and by the determined video recommendations that will recommend to user, comprising:
According to the height of the interest value of each video features, select the high interest value video features of the 3rd number;
For the high interest value video features of the 3rd selected number, from video library, select the video of the 4th number, wherein, each video in the video of described 4th number has at least one high interest value video features;
Calculate the comprehensive interest value of each video in the video of described 4th number, described comprehensive interest value be the interest value of all high interest value video features of this video and;
According to the height of described comprehensive interest value, recommend at least one video in the video of described 4th number to user.
The embodiment of the invention also discloses a kind of video recommendations device, technical scheme is as follows:
A kind of video recommendations device, be applied to server, store the mark of video and the video features corresponding with the mark of video in described server this locality, described device comprises:
User behavior acquiring unit, for obtain user for the operation behavior of video and described operation behavior for the mark of video;
Video features acquiring unit, for extracting the multiple video features corresponding with the mark of obtained video from this locality;
Interest value computing unit, for being all handled as follows each extracted video features: according to the behavior weight corresponding with this video features characteristic of correspondence weight and the default described operation behavior with obtaining preset, calculate the interest value obtaining this video features;
Storage unit, for by the interest value of obtained each video features and user profile corresponding stored;
Recommendation unit, for the video determining to recommend according to the interest value of stored each video features corresponding with user profile and by the determined video recommendations that will recommend to user.
Preferably, described interest value computing unit comprises:
Weight determination subelement, for determining the default feature weight corresponding with this video features and the default behavior weight corresponding with described operation behavior, wherein, all default with one the behavior weight of each operation behavior is corresponding;
Interest value computation subunit, for when obtained user is one for the operation behavior of video, by determined default feature weight and determined default behavior multiplied by weight, obtains the interest value of this video features; When obtained user is multiple for the operation behavior of video, by determined default feature weight respectively with determined each default behavior multiplied by weight, by each product addition, obtain the interest value of this video features.
Preferably, described storage unit specifically for: by the interest value of obtained each video features and the rise time of each interest value and user profile corresponding stored;
Described recommendation unit comprises:
Interest value weighting process subelement, for the rise time according to each interest value, is weighted process to each interest value;
First recommends subelement, for the video determining to recommend according to the interest value after weighting process and by the determined video recommendations that will recommend to user.
Preferably, described recommendation unit comprises:
First high interest value video features chooser unit, for the height of the interest value according to each video features, selects the high interest value video features of the first number;
First video chooser unit, for the high interest value video features for the first selected number, selects the video of the second number from video library, and wherein, each video in the video of described second number has at least one high interest value video features;
Second recommends subelement, for being recommended to user by the video of described second number.
Preferably, described recommendation unit comprises:
Second high interest value video features chooser unit, for the height of the interest value according to each video features, selects the high interest value video features of the 3rd number;
Second video chooser unit, for the high interest value video features for the 3rd selected number, from video library, select the video of the 4th number, wherein, each video in the video of described 4th number has at least one high interest value video features;
Comprehensive interest value computation subunit, for calculating the comprehensive interest value of each video in the video of described 4th number, described comprehensive interest value be the interest value of all high interest value video features of this video and;
3rd recommends subelement, for the height according to described comprehensive interest value, recommends at least one video in the video of described 4th number to user.
A kind of video recommendation method that the embodiment of the present invention provides and device, obtain the operation behavior of user and the mark of operated video, to each video features corresponding to this video labeling, the interest value of user to this video features is calculated based on information such as obtained operation behaviors, by interest value and user profile corresponding stored, and recommend video according to stored interest value to this user.Because the present invention is when recommending video to user, consider this user operation behavior in the past and the feature of operated video, thus recommended video more meets user interest, improves the effect of video recommendations.
Accompanying drawing explanation
In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is the schematic flow sheet of a kind of video recommendation method of the embodiment of the present invention;
Fig. 2 is the structural representation of a kind of video recommendations device of the embodiment of the present invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.
In order to solve prior art problem, embodiments provide the various signature analysis user interests of a kind of concrete behavior details from user operation video and operated video, and and then recommend the method and apparatus of video according to interest, improve the effect of video recommendations.
See Fig. 1, Figure 1 shows that a kind of schematic flow sheet of a kind of video recommendation method of the embodiment of the present invention, can comprise the following steps:
S101, obtain user for the operation behavior of video and described operation behavior for the mark of video;
After video website detects that a certain user has carried out viewing or other operations to a certain video, this step can obtain operation behavior information and the video labeling of user, namely which video user have viewed, and whether user has carried out complete broadcasting, have viewed and how long, whether to have carried out scoring, commented how many points, whether put information such as praising.
S102, extracts the multiple video features corresponding with the mark of obtained video from this locality;
According to the video labeling that step S101 obtains, extract the various video feature corresponding to this video labeling.Video features can divide into the video features of multiple different type, such as: video regional feature, video time feature, video story of a play or opera feature, video duration characteristics, video performers and clerks feature etc.Video regional feature can comprise: " Hong Kong and Taiwan are acute " " day South Korean TV soaps " " continent is acute " " American-European acute " etc., video story of a play or opera feature can comprise: " imperial palace is acute " " ancient costume acrobatic fighting is acute " " modern romantic play " etc., and video time feature can comprise: " the eighties ", " the nineties ", " newly acute " etc.Because video features is distinguished thinner, the user interest that final analysis of the present invention goes out will be more careful, and correspondingly, the video recommendations effect done eventually to user can be better, therefore, when calculating and the storage capacity permission of server, more detailed differentiation can be carried out to the feature of video.For example, if user have viewed " discriminating Huan to pass " the 2nd collection, this video meets " imperial palace is acute " and " continent is acute " these two video features in above-mentioned classification simultaneously.In subsequent step, need the calculating " imperial palace is acute " and " continent is acute " these two kinds of features being carried out to interesting data.
S103, is all handled as follows each extracted video features: according to the behavior weight corresponding with this video features characteristic of correspondence weight and the default described operation behavior with obtaining preset, calculate the interest value obtaining this video features;
The object of this step is to calculate interesting data.So-called calculating interesting data, that is: be directed to each feature that step S102 extracts, and calculates user to the interest of this feature.In embodiments of the present invention, interest represents with interest value, and interest value is a numeral being greater than zero.
Corresponding to each video features, a weight is given in the invention process regular meeting in advance.These weight coefficients can be equal or not etc.In general, the video features that broad scope is more wide in range, such as " American-European acute ", its feature weight can be established lower, and the feature that broad scope is narrower, such as " ancient costume acrobatic fighting is acute ", its weight can be established higher, this is because the video features that broad scope is more wide in range, from user, its operation is shown that the confidence level of user to its interesting conclusion is lower.
The embodiment of the present invention is also distinguished different user operation behaviors and is that each operation behavior gives a behavior weight in advance.Such as, can in advance operation behavior be divided into: " viewing integrity degree ", " download ", " scoring ", " point is praised ", " comment " etc.Corresponding to each operation behavior, the embodiment of the present invention gives a weight coefficient in advance, in subsequent step to the calculating of interesting data.
Weight coefficient corresponding to a certain operation behavior is set to how many, depends on for analyzing the object of interesting data, this operation behavior can with reference to degree.Such as can give higher weight, because download behavior often represents higher interest to the download behavior of user.And user may have multiple to the operation behavior of a video, then need to consider these multiple operation behaviors when calculating interest value.
Obviously more than a kind of according to the method for video features and feature weight, operation behavior and behavior weight calculation interest value, wherein a kind of method compared with simple and feasible can be:
Determine the default feature weight corresponding with this video features and the default behavior weight corresponding with described operation behavior, wherein, all default with one the behavior weight of each operation behavior is corresponding; When obtained user is one for the operation behavior of video, by determined default feature weight and determined default behavior multiplied by weight, obtain the interest value of this video features; When obtained user is multiple for the operation behavior of video, by determined default feature weight respectively with determined each default behavior multiplied by weight, by each product addition, obtain the interest of this video features.
Illustrate below.If a certain user have viewed, " discriminating Huan to pass " the 2nd collects, this video will meet " imperial palace is acute " and " continent is acute " these two video features simultaneously, then need for each feature calculation interest value in " imperial palace is acute " and " continent is acute " these two video features.Suppose that this user has carried out complete viewing and put praising to Dream of the Red Mansion the 2nd collection, then, when calculating each interest value, need to consider " viewing integrity degree " and " point is praised " these two operation behaviors.
Suppose that " imperial palace is acute " and the default feature weight corresponding to " continent is acute " are respectively 0.4 and 0.2, the default behavior weight corresponding to " viewing integrity degree " and " point is praised " is respectively 0.3 and 0.6.Understandable, why " imperial palace is acute " being arranged higher weight relative to " continent is acute ", is because imperial palace play is a video features that broad scope is narrower comparatively speaking; Why " point is praised " being arranged higher weight relative to " viewing integrity degree ", is more can representative of consumer be interested in this video because obviously put behavior of praising comparatively speaking.
Therefore, calculating acquisition user for the interest value of " imperial palace is acute " is: 0.4*0.3+0.4*0.6=0.36; User is 0.2*0.3+0.2*0.6=0.18 for the interest value of " continent is acute ".
Lift an example again.Suppose: user have viewed " En Ter The Matrix " first, this video meets simultaneously " science fiction acute " and " American-European play " these two video features; This user has carried out complete viewing to this video, has put and praised and download; " science fiction is acute " that preset and " American-European acute " characteristic of correspondence weight is respectively 0.4 and 0.2, and " the viewing integrity degree ", " point is praised " that preset, the behavior weight corresponding to " download " are respectively 0.3,0.6 and 0.9.
The result of calculation of each interest value is as follows: user for the interest value of " science fiction is acute " is: 0.4*0.3+0.4*0.6+0.4*0.9=0.72, and user is 0.2*0.3+0.2*0.6+0.2*0.9=0.36 for the interest value of " American-European acute ".
The interest value that this example calculates wants high relative to the interest value calculated in " discriminating Huan to pass " example, main cause is exactly had more " download " this kind of behavior in user behavior, this embodies higher user interest, so final interest value data are also general higher.
S104, by the interest value of obtained each video features and user profile corresponding stored;
After step S103 calculating acquisition user is to the interest value of multiple video features, by these interest value and user profile corresponding stored, recommend video for this this user backward.
Step S103 is calculated to the interest value obtained, can directly store.Certainly, in another embodiment of the present invention, also can detect the new video features corresponding to interest value obtained that calculates and whether there is the interest value stored on the server, if existed, accumulation interest value is determined with the interest value stored according to the new interest value obtained that calculates, after this, determined accumulation interest value and accumulation interest value are fixed time and user profile corresponding stored really.
For example, the time that same user may be separated by front and back, more than once viewed " science fiction is acute ".The viewing each time of user, all can produce the interest value of this user for " science fiction is acute ".Therefore, after obtaining the new interest value to " science fiction is acute ", the embodiment of the present invention determines accumulation interest value by calculating according to the new interest value (hereinafter referred to as new interest value) calculating acquisition and the interest value (hereinafter referred to as old interest value) stored, and by accumulation interest value and user profile corresponding stored, the accumulation interest value that also has stored is fixed time really simultaneously.
The computing method of accumulation interest value can have many, for example, old interest value and new interest value can be multiplied by a weight coefficient respectively, then product summation are obtained.Ratio as usual interest value is 0.5, and new interest value is 0.72, then accumulation interest value is w1*0.5+w2*0.75.Wherein w1+w2=1, w1<w2.Such computing method are to give new interest value with higher weight when calculating accumulation interest value.
When calculate obtain accumulation interest value, the various operations done according to interest value mentioned in the embodiment of the present invention also can be understood as the various operations done according to accumulation interest value natch.
In another preferred embodiment, consider that the interest of user may occur " migration " along with the time, at the video determining to recommend according to the interest value of stored each video features corresponding with user profile and by the determined video recommendations that will recommend to the step of user before, can also comprise: by the rise time of the interest value of obtained each video features and user profile corresponding stored.
The object done like this is when recommending according to interest value in subsequent step S105, consider rise time more late interest value more and less consider the rise time comparatively early interest value.
To sum up, the execution object of this step is the interest value of acquisition to store, storage mentioned here is included in and finds that same video features has stored when haveing been friends in the past interest value, obtains accumulation interest value and the accumulation interest value of this video features stored according to new and old interest value by weighted calculation.
It is emphasized that above all steps, from S101 to S104, can be when server processing power allow, immediately obtain perform.This instantaneity can ensure the best ageing of the recommendation done in subsequent step S105.Some embodiment of the present invention adopts the processing mode of increment streaming, once detect that user has carried out operation behavior, then calculates and/or upgrade the interest value data of this user immediately.This instantaneity is the important preferred technique effect of the embodiment of the present invention.
S105, determines according to the interest value of stored each video features corresponding with user profile the video that will recommend and by the determined video recommendations that will recommend to user.
In this step, first read the interest value stored, find out the video features that some interest value are higher, from video library, then select some videos with these high interest value video features recommend to user.
A preferred concrete executive mode of S105 can be: according to the height of the interest value of each video features, select the high interest value video features of the first number;
For the high interest value video features of the first selected number, select the video of the second number from video library, wherein, each video in the video of described second number has at least one high interest value video features;
The video of described second number is recommended to user.
For example, if want to recommend 4 videos to user, a kind of implementation method is, first find out from each interest value of this user stored and be no more than 4 interest value the highest, each in these high interest value correspond to a high interest value video features, then from video library, select 4 videos recommend to user, wherein each video possesses at least one high interest value video features.
Indicated by above-mentioned example is an embodiment more simply.Many better recommend methods can also be had.
Another preferred concrete executive mode of S105 can be: according to the height of the interest value of each video features, select the high interest value video features of the 3rd number;
For the high interest value video features of the 3rd selected number, from video library, select the video of the 4th number, wherein, each video in the video of described 4th number has at least one high interest value video features;
Calculate the comprehensive interest value of each video in the video of described 4th number, described comprehensive interest value be the interest value of all high interest value video features of this video and;
According to the height of described comprehensive interest value, recommend at least one video in the video of described 4th number to user.
Such as want to recommend 4 videos to user equally, a kind of more excellent method is, finds out more than 4 from each interest value of this user, such as 10 high interest value, these 10 high interest value correspond to 10 high interest value video features, then continuous Picking video from video library.Consider that a video may have a more than high interest value video features, so when Picking video, summation operation is carried out to the high interest value video features that each video has, obtains the comprehensive interest value of this video.Under this implementation, 4 videos recommended eventually to user are exactly the video possessing higher comprehensive interest value picked out in video library.Relative to upper a kind of plain mode, the benefit of the manner is just to consider a kind of situation, namely a certain video may not have an extra high video features of interest value, but it has the higher video features of several interest value simultaneously, so such video also should be selected in and recommend to user.
In this step, to each interest value stored, can directly read and use.
Another preferred concrete executive mode of this step can be: according to the rise time of each interest value, be weighted process to each interest value;
When interest value is accumulation interest value, this rise time can determine the time for the last calculating of accumulation interest value.
The video that will recommend is determined and by the determined video recommendations that will recommend to user according to the interest value after weighting process.
After reading the interest value stored, be used further to after interest value is carried out being weighted process according to the rise time as recommendation foundation.Concrete grammar can be the interest value for more early producing, and is just multiplied by a less coefficient, for the interest value of more late generation, is just multiplied by a larger coefficient, obtains the interest value after the temporally process of weighting sooner or later.
For the interest value that the example in step S103 obtains, then store altogether four interest value, they respectively:
The interest value of " imperial palace is acute " is 0.4*0.3+0.4*0.6=0.36;
The interest value of " continent is acute " is 0.2*0.3+0.2*0.6=0.18;
The interest value of " science fiction is acute " is 0.4*0.3+0.4*0.6+0.4*0.9=0.72;
The interest value of " American-European acute " is 0.2*0.3+0.2*0.6+0.2*0.9=0.36;
Front two articles of interest value calculate according to user's operation behavior to " discriminating Huan biography " the 2nd collection to obtain, and latter two is calculate acquisition according to user to the operation behavior of " En Ter The Matrix " first.But before supposing that the operation behavior of user to Dream of the Red Mansion occurs in half a year, and before three days are occurred in the operation behavior of " En Ter The Matrix ", then consider time factor, a weighting coefficient 0.5 can be multiplied by the interest value obtained because of user operation Dream of the Red Mansion, and to the interest value obtained because of user operation " En Ter The Matrix ", because before occurring over just three days, be then multiplied by weighting coefficient 1.Here, weighting coefficient 0.5 or 1 is all only and illustrates, in actual applications, can be set to other and reasonably be worth.
So the interest value after weighting process becomes: the interest value of " imperial palace is acute " is 0.18; The interest value of " continent is acute " is 0.09; The interest value of " science fiction is acute " is 0.72; The interest value of " American-European acute " is 0.36.Originally equal with " American-European acute " interest value " imperial palace is acute ", because the time is remote, interest value reduces half, then likely just can not become the foundation of video recommendations.
Embodiment is as shown in Figure 1 visible, this video recommendation method that the embodiment of the present invention provides, from user watch video concrete behavior details and watch the various signature analysis user interests of video, and in some embodiments, it is also conceivable to the new and old of user interest, and and then according to interest recommend video, compared with prior art significantly improve the effect of video recommendations.
Corresponding to embodiment of the method above, present invention also offers one video recommendations device.
See Fig. 2, Figure 2 shows that the structural representation of a kind of video recommendations device that the embodiment of the present invention provides.This application of installation is in server, store the mark of video and the video features corresponding with the mark of video in described server this locality, comprising: user behavior acquiring unit 201, video features acquiring unit 202, interest value computing unit 203, storage unit 204 and recommendation unit 205.Wherein:
User behavior acquiring unit 201, for obtain user for the operation behavior of video and described operation behavior for the mark of video;
Video features acquiring unit 202, for extracting the multiple video features corresponding with the mark of obtained video from this locality;
Interest value computing unit 203, for being all handled as follows each extracted video features: according to the behavior weight corresponding with this video features characteristic of correspondence weight and the default described operation behavior with obtaining preset, calculate the interest value obtaining this video features;
In a preferred embodiment, interest value computing unit 203 comprises:
Weight determination subelement, for determining the default feature weight corresponding with this video features and the default behavior weight corresponding with described operation behavior, wherein, all default with one the behavior weight of each operation behavior is corresponding;
Interest value computation subunit, for when obtained user is one for the operation behavior of video, by determined default feature weight and determined default behavior multiplied by weight, obtains the interest value of this video features; When obtained user is multiple for the operation behavior of video, by determined default feature weight respectively with determined each default behavior multiplied by weight, by each product addition, obtain the interest value of this video features.
Storage unit 204, for by the interest value of obtained each video features and user profile corresponding stored;
In a preferred embodiment, consider that the interest of user may occur " migration " along with the time, storage unit 204 specifically for: by the interest value of obtained each video features and the rise time of each interest value and user profile corresponding stored.
Recommendation unit 205 specifically for: for the video determining to recommend according to the interest value of stored each video features corresponding with user profile and by the determined video recommendations that will recommend to user.
In a preferred embodiment, recommendation unit 205 can comprise:
First high interest value video features chooser unit, for the height of the interest value according to each video features, selects the high interest value video features of the first number;
First video chooser unit, for the high interest value video features for the first selected number, selects the video of the second number from video library, and wherein, each video in the video of described second number has at least one high interest value video features;
Second recommends subelement, for being recommended to user by the video of described second number.
In a preferred embodiment, recommendation unit 205 can also comprise:
Second high interest value video features chooser unit, for the height of the interest value according to each video features, selects the high interest value video features of the 3rd number;
Second video chooser unit, for the high interest value video features for the 3rd selected number, from video library, select the video of the 4th number, wherein, each video in the video of described 4th number has at least one high interest value video features;
Comprehensive interest value computation subunit, for calculating the comprehensive interest value of each video in the video of described 4th number, described comprehensive interest value be the interest value of all high interest value video features of this video and;
3rd recommends subelement, for the height according to described comprehensive interest value, recommends at least one video in the video of described 4th number to user.
Two kinds of preferred embodiments are compared, the benefit of the latter is to consider a kind of situation, and namely a certain video may not have an extra high video features of interest value, but it has the higher video features of several interest value simultaneously, so such video also can be selected and recommend to user.
When storage unit 204 is specifically for during by the interest value of obtained each video features and the rise time of each interest value and user profile corresponding stored, described recommendation unit 205 can comprise:
Interest value weighting process subelement, for the rise time according to each interest value, is weighted process to each interest value;
First recommends subelement, for the video determining to recommend according to the interest value after weighting process and by the determined video recommendations that will recommend to user.
Embodiment is as shown in Figure 2 visible, this video recommendations device that the embodiment of the present invention provides, from user watch video concrete behavior details and watch the various signature analysis user interests of video, and consider the new and old of user interest, and and then according to interest recommend video, compared with prior art significantly improve the effect of video recommendations.
Those skilled in the art can be well understood to, for convenience and simplicity of description, the device of foregoing description and the specific works process of module, can describe with reference to the corresponding process in preceding method embodiment, therefore, in this device illustrates, describe simpler, only some focus technology main points are described.
It should be noted that, in this article, the such as relational terms of first and second grades and so on is only used for an entity or operation to separate with another entity or operational zone, and not necessarily requires or imply the relation that there is any this reality between these entities or operation or sequentially.And, term " comprises ", " comprising " or its any other variant are intended to contain comprising of nonexcludability, thus make to comprise the process of a series of key element, method, article or equipment and not only comprise those key elements, but also comprise other key elements clearly do not listed, or also comprise by the intrinsic key element of this process, method, article or equipment.When not more restrictions, the key element limited by statement " comprising ... ", and be not precluded within process, method, article or the equipment comprising described key element and also there is other identical element.
Each embodiment in this instructions all adopts relevant mode to describe, between each embodiment identical similar part mutually see, what each embodiment stressed is the difference with other embodiments.Especially, for device embodiment, because it is substantially similar to embodiment of the method, so description is fairly simple, relevant part illustrates see the part of embodiment of the method.
One of ordinary skill in the art will appreciate that all or part of step realized in said method embodiment is that the hardware that can carry out instruction relevant by program has come, described program can be stored in computer read/write memory medium, here the alleged storage medium obtained, as: ROM/RAM, magnetic disc, CD etc.
The foregoing is only preferred embodiment of the present invention, be not intended to limit protection scope of the present invention.All any amendments done within the spirit and principles in the present invention, equivalent replacement, improvement etc., be all included in protection scope of the present invention.
Claims (10)
1. a video recommendation method, is characterized in that, is applied to server, and store the mark of video and the video features corresponding with the mark of video in described server this locality, described method comprises:
Obtain user for the operation behavior of video and described operation behavior for the mark of video;
The multiple video features corresponding with the mark of obtained video are extracted from this locality;
Each extracted video features is all handled as follows: according to the behavior weight corresponding with this video features characteristic of correspondence weight and the default described operation behavior with obtaining preset, calculate the interest value obtaining this video features;
By the interest value of obtained each video features and user profile corresponding stored;
The video that will recommend is determined and by the determined video recommendations that will recommend to user according to the interest value of stored each video features corresponding with user profile.
2. method according to claim 1, is characterized in that, described basis preset with this video features characteristic of correspondence weight and the default behavior weight corresponding with obtained operation behavior, calculate and obtain the interest value of this video features, comprising:
Determine the default feature weight corresponding with this video features and the default behavior weight corresponding with described operation behavior, wherein, all default with one the behavior weight of each operation behavior is corresponding; When obtained user is one for the operation behavior of video, by determined default feature weight and determined default behavior multiplied by weight, obtain the interest value of this video features; When obtained user is multiple for the operation behavior of video, by determined default feature weight respectively with determined each default behavior multiplied by weight, by each product addition, obtain the interest value of this video features.
3. method according to claim 1, it is characterized in that, the described interest value according to the stored each video features corresponding with user profile is determined the video that will recommend and by the determined video recommendations that will recommend to before the step of user, is also comprised: by the rise time of the interest value of obtained each video features and user profile corresponding stored;
The described interest value according to the stored each video features corresponding with user profile is determined the video that will recommend and by the determined video recommendations that will recommend to user, being comprised:
According to the rise time of each interest value, process is weighted to each interest value;
The video that will recommend is determined and by the determined video recommendations that will recommend to user according to the interest value after weighting process.
4. method according to claim 1, is characterized in that, described to determine recommending with the interest value of each video features of user's corresponding stored according to stored video and by the determined video recommendations that will recommend to user, comprising:
According to the height of the interest value of each video features, select the high interest value video features of the first number;
For the high interest value video features of the first selected number, select the video of the second number from video library, wherein, each video in the video of described second number has at least one high interest value video features;
The video of described second number is recommended to user.
5. method according to claim 1, is characterized in that, described to determine recommending with the interest value of each video features of user's corresponding stored according to stored video and by the determined video recommendations that will recommend to user, comprising:
According to the height of the interest value of each video features, select the high interest value video features of the 3rd number;
For the high interest value video features of the 3rd selected number, from video library, select the video of the 4th number, wherein, each video in the video of described 4th number has at least one high interest value video features;
Calculate the comprehensive interest value of each video in the video of described 4th number, described comprehensive interest value be the interest value of all high interest value video features of this video and;
According to the height of described comprehensive interest value, recommend at least one video in the video of described 4th number to user.
6. a video recommendations device, is characterized in that, is applied to server, and store the mark of video and the video features corresponding with the mark of video in described server this locality, described device comprises:
User behavior acquiring unit, for obtain user for the operation behavior of video and described operation behavior for the mark of video;
Video features acquiring unit, for extracting the multiple video features corresponding with the mark of obtained video from this locality;
Interest value computing unit, for being all handled as follows each extracted video features: according to the behavior weight corresponding with this video features characteristic of correspondence weight and the default described operation behavior with obtaining preset, calculate the interest value obtaining this video features;
Storage unit, for by the interest value of obtained each video features and user profile corresponding stored;
Recommendation unit, for the video determining to recommend according to the interest value of stored each video features corresponding with user profile and by the determined video recommendations that will recommend to user.
7. device according to claim 6, is characterized in that, described interest value computing unit comprises:
Weight determination subelement, for determining the default feature weight corresponding with this video features and the default behavior weight corresponding with described operation behavior, wherein, all default with one the behavior weight of each operation behavior is corresponding;
Interest value computation subunit, for when obtained user is one for the operation behavior of video, by determined default feature weight and determined default behavior multiplied by weight, obtains the interest value of this video features; When obtained user is multiple for the operation behavior of video, by determined default feature weight respectively with determined each default behavior multiplied by weight, by each product addition, obtain the interest value of this video features.
8. device according to claim 6, is characterized in that,
Described storage unit specifically for: by the interest value of obtained each video features and the rise time of each interest value and user profile corresponding stored;
Described recommendation unit comprises:
Interest value weighting process subelement, for the rise time according to each interest value, is weighted process to each interest value;
First recommends subelement, for the video determining to recommend according to the interest value after weighting process and by the determined video recommendations that will recommend to user.
9. device according to claim 6, is characterized in that, described recommendation unit comprises:
First high interest value video features chooser unit, for the height of the interest value according to each video features, selects the high interest value video features of the first number;
First video chooser unit, for the high interest value video features for the first selected number, selects the video of the second number from video library, and wherein, each video in the video of described second number has at least one high interest value video features;
Second recommends subelement, for being recommended to user by the video of described second number.
10. device according to claim 6, is characterized in that, described recommendation unit comprises:
Second high interest value video features chooser unit, for the height of the interest value according to each video features, selects the high interest value video features of the 3rd number;
Second video chooser unit, for the high interest value video features for the 3rd selected number, from video library, select the video of the 4th number, wherein, each video in the video of described 4th number has at least one high interest value video features;
Comprehensive interest value computation subunit, for calculating the comprehensive interest value of each video in the video of described 4th number, described comprehensive interest value be the interest value of all high interest value video features of this video and;
3rd recommends subelement, for the height according to described comprehensive interest value, recommends at least one video in the video of described 4th number to user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510150095.XA CN104731950B (en) | 2015-03-31 | 2015-03-31 | A kind of video recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510150095.XA CN104731950B (en) | 2015-03-31 | 2015-03-31 | A kind of video recommendation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104731950A true CN104731950A (en) | 2015-06-24 |
| CN104731950B CN104731950B (en) | 2018-09-28 |
Family
ID=53455837
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510150095.XA Active CN104731950B (en) | 2015-03-31 | 2015-03-31 | A kind of video recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104731950B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095431A (en) * | 2015-07-22 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Method and device for pushing videos based on behavior information of user |
| CN105512331A (en) * | 2015-12-28 | 2016-04-20 | 海信集团有限公司 | Video recommending method and device |
| CN105843953A (en) * | 2016-04-12 | 2016-08-10 | 乐视控股（北京）有限公司 | Multimedia recommendation method and device |
| CN105872593A (en) * | 2016-03-21 | 2016-08-17 | 乐视网信息技术（北京）股份有限公司 | Barrage pushing method and device |
| CN105898410A (en) * | 2015-12-15 | 2016-08-24 | 乐视网信息技术（北京）股份有限公司 | Video recommendation method and server |
| CN106507143A (en) * | 2016-10-21 | 2017-03-15 | 北京小米移动软件有限公司 | Video recommendation method and device |
| CN106792172A (en) * | 2016-12-14 | 2017-05-31 | 未来电视有限公司 | A kind of method of internet television personalized recommendation video |
| CN106845730A (en) * | 2017-02-16 | 2017-06-13 | 深圳大学 | A kind of algorithm of credit rating containing when based on scoring network |
| CN107748759A (en) * | 2017-09-21 | 2018-03-02 | 广州华多网络科技有限公司 | Video pushing method and its equipment |
| CN107844598A (en) * | 2017-11-22 | 2018-03-27 | 广州优视网络科技有限公司 | Content recommendation method, device and computer equipment |
| CN107968958A (en) * | 2016-10-20 | 2018-04-27 | 法乐第(北京)网络科技有限公司 | Video recommendation method, device and terminal device |
| CN108170868A (en) * | 2018-02-09 | 2018-06-15 | 宁夏灵智科技有限公司 | Video recommendation method and device |
| WO2018148894A1 (en) * | 2017-02-16 | 2018-08-23 | 深圳大学 | Time-dependent reputation evaluation algorithm based on scoring network |
| CN110096613A (en) * | 2019-04-12 | 2019-08-06 | 北京奇艺世纪科技有限公司 | A kind of video recommendation method, device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102946566A (en) * | 2012-10-24 | 2013-02-27 | 北京奇虎科技有限公司 | Video recommending method and device based on historical information |
| CN103870454A (en) * | 2012-12-07 | 2014-06-18 | 盛乐信息技术（上海）有限公司 | Method and method for recommending data |
| CN104216960A (en) * | 2014-08-21 | 2014-12-17 | 北京奇艺世纪科技有限公司 | Method and device for recommending video |
-
2015
- 2015-03-31 CN CN201510150095.XA patent/CN104731950B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102946566A (en) * | 2012-10-24 | 2013-02-27 | 北京奇虎科技有限公司 | Video recommending method and device based on historical information |
| CN103870454A (en) * | 2012-12-07 | 2014-06-18 | 盛乐信息技术（上海）有限公司 | Method and method for recommending data |
| CN104216960A (en) * | 2014-08-21 | 2014-12-17 | 北京奇艺世纪科技有限公司 | Method and device for recommending video |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095431A (en) * | 2015-07-22 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Method and device for pushing videos based on behavior information of user |
| CN105898410A (en) * | 2015-12-15 | 2016-08-24 | 乐视网信息技术（北京）股份有限公司 | Video recommendation method and server |
| CN105512331A (en) * | 2015-12-28 | 2016-04-20 | 海信集团有限公司 | Video recommending method and device |
| CN105512331B (en) * | 2015-12-28 | 2019-03-26 | 海信集团有限公司 | A kind of video recommendation method and device |
| CN105872593A (en) * | 2016-03-21 | 2016-08-17 | 乐视网信息技术（北京）股份有限公司 | Barrage pushing method and device |
| WO2017177643A1 (en) * | 2016-04-12 | 2017-10-19 | 乐视控股（北京）有限公司 | Multimedia recommendation method and device |
| CN105843953A (en) * | 2016-04-12 | 2016-08-10 | 乐视控股（北京）有限公司 | Multimedia recommendation method and device |
| CN107968958A (en) * | 2016-10-20 | 2018-04-27 | 法乐第(北京)网络科技有限公司 | Video recommendation method, device and terminal device |
| CN106507143A (en) * | 2016-10-21 | 2017-03-15 | 北京小米移动软件有限公司 | Video recommendation method and device |
| CN106792172A (en) * | 2016-12-14 | 2017-05-31 | 未来电视有限公司 | A kind of method of internet television personalized recommendation video |
| CN106845730A (en) * | 2017-02-16 | 2017-06-13 | 深圳大学 | A kind of algorithm of credit rating containing when based on scoring network |
| WO2018148894A1 (en) * | 2017-02-16 | 2018-08-23 | 深圳大学 | Time-dependent reputation evaluation algorithm based on scoring network |
| CN106845730B (en) * | 2017-02-16 | 2021-05-18 | 深圳大学 | Time-containing reputation evaluation method based on scoring network |
| CN107748759A (en) * | 2017-09-21 | 2018-03-02 | 广州华多网络科技有限公司 | Video pushing method and its equipment |
| CN107844598A (en) * | 2017-11-22 | 2018-03-27 | 广州优视网络科技有限公司 | Content recommendation method, device and computer equipment |
| CN108170868A (en) * | 2018-02-09 | 2018-06-15 | 宁夏灵智科技有限公司 | Video recommendation method and device |
| CN110096613A (en) * | 2019-04-12 | 2019-08-06 | 北京奇艺世纪科技有限公司 | A kind of video recommendation method, device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104731950B (en) | 2018-09-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104731950A (en) | Video recommendation method and device | |
| CN110704674B (en) | Video playing integrity prediction method and device | |
| KR101729195B1 (en) | System and Method for Searching Choreography Database based on Motion Inquiry | |
| CN101840410B (en) | Learning apparatus, learning method, information processing apparatus, method and program | |
| JP7493323B2 (en) | Information processing device, method for controlling information processing device, and program | |
| CN106599165B (en) | content recommendation method and server based on playing behavior | |
| CN105847985A (en) | Video recommendation method and device | |
| CN105095431A (en) | Method and device for pushing videos based on behavior information of user | |
| CN109753601B (en) | Method and device for determining click rate of recommended information and electronic equipment | |
| CN101281540A (en) | Apparatus, method and computer program for processing information | |
| CN104504059A (en) | Multimedia resource recommending method | |
| CN105915956A (en) | Video content recommendation method, device, server and system | |
| KR101804170B1 (en) | Item recommendation method and apparatus thereof utilizing uninteresting item and apparatus | |
| CN105718545A (en) | Recommendation method and device of multimedia resources | |
| CN106599047B (en) | Information pushing method and device | |
| CN110046278B (en) | Video classification method and device, terminal equipment and storage medium | |
| CN105308636A (en) | Apparatus and method for providing virtual advertisement | |
| CN104008139A (en) | Method and device for creating video index table and method and device for recommending video | |
| CN106033455B (en) | Method and equipment for processing user operation information | |
| CN113407773A (en) | Short video intelligent recommendation method and system, electronic device and storage medium | |
| CN106227870A (en) | A kind of video website recommends method and device | |
| CN109379608A (en) | A kind of recommended method and relevant device of direct broadcasting room | |
| US20130108180A1 (en) | Information processing device, information processing method, and program | |
| CN111046230A (en) | Content recommendation method and device, electronic equipment and storable medium | |
| CN109063080B (en) | Video recommendation method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |