Determine the method and device of the title of multimedia resource
Technical field
The present invention relates to areas of information technology, the method and device of more particularly, to a kind of title determining multimedia resource.
Background technology
In Internet era, especially mobile Internet epoch, how to provide the user timely and valuable information is
The focus of numerous Internet firms research.For example, user is when browsing video website, and video title is to attract user's viewing video
A key factor, therefore, video website often has substantial amounts of operation personnel to enter edlin to video title.Video uploader
Edlin can also be entered to video title, to reach the purpose attracting user's viewing.
At present, the editor of the title of the multimedia resource such as video depends on operation personnel and the uploader of website, expends big
The human resources of amount, and the title of multimedia resource that the operation personnel of website and uploader are edited is for popular hobby
It is impossible to meet the individual demand of unique user.
Content of the invention
Technical problem
In view of this, the technical problem to be solved in the present invention is, the mode of the existing title determining multimedia resource consumes
Take substantial amounts of human resources, and the individual demand of user can not be met.
Solution
In order to solve above-mentioned technical problem, according to one embodiment of the invention, there is provided a kind of determination multimedia resource
The method of title, including:
The user behavior data of collection targeted customer, and the first multimedia resource row are generated according to described user behavior data
Table;
The title of each multimedia resource in described first multimedia resource list is parsed, obtains described target and use
Family corresponding first composition list;
The former title of multimedia resource to be recommended is parsed, obtains the corresponding second composition list of described former title;
Each composition in described second composition list is compared with each composition in described first composition list,
Second composition list after being updated;
Determine the new title of described multimedia resource to be recommended according to the second composition list after described renewal.
For said method, in a kind of possible implementation, by each composition in described second composition list with
Each composition in described first composition list is compared, the second composition list after being updated, including:
Each composition calculating in described second composition list is similar to each composition in described first composition list
Degree;
The similarity of the composition in the composition in described second composition list with described first composition list is more than
In the case of first preset value, the composition in described second composition list is replaced with an one-tenth in described first composition list
Point;
Second composition list after being updated according to the composition of all replacements.
For said method, in a kind of possible implementation, calculate each composition in described second composition list
With the similarity of each composition in described first composition list, including:
Determine the corresponding vector of each composition in described second composition list;
Calculate respectively each composition in described second composition list corresponding vectorial with described first composition list in
The similarity of the corresponding vector of each composition.
For said method, in a kind of possible implementation, calculate in described second composition list respectively each
Composition is corresponding vectorial and described first composition list in the corresponding vector of each composition similarity, including:
Calculate the corresponding vector of l-th composition in described second composition list using formula 1With described first composition list
In the corresponding vector of m-th compositionSimilarity
For said method, in a kind of possible implementation, determined according to the second composition list after described renewal
The new title of described multimedia resource to be recommended, including:
Calculate the score of the second composition list after described renewal;
In the case that the score of the second composition list after described renewal is more than the second preset value, after described renewal
Second composition list determine the new title of described multimedia resource to be recommended.
For said method, in a kind of possible implementation, calculate obtaining of the second composition list after described renewal
Point, including:
The probability meter being occurred in specified sample set according to each composition in the second composition list after described renewal
Calculate the score of the second composition list after described renewal.
For said method, in a kind of possible implementation, according in the second composition list after described renewal
The score of the second composition list after updating described in the probability calculation that each composition occurs in specified sample set, including:
Calculate score s of the second composition list after described renewal using formula 2;
Wherein, the number of composition, w in the second composition list after n represents described renewaljRepresent second after described renewal
J-th composition in ingredient lists, wj-iRepresent jth-i composition in the second composition list after described renewal, p (wjwj-i)
Represent described j-th composition and described jth-i the composition common probability occurring, p (w in described specified sample setj-i) table
Show the probability that described jth-i composition occurs in described specified sample set.
A kind of second composition list for said method, in possible implementation, after calculating described renewal
After score, methods described also includes:
In the case that the score of the second composition list after described renewal is less than or equal to described second preset value, retain
The former title of described multimedia resource to be recommended.
For said method, in a kind of possible implementation, in described first multimedia resource list, each is many
The title of media resource is parsed, and obtains described targeted customer corresponding first composition list, including:
The title of each multimedia resource in described first multimedia resource list is parsed, obtains and described target
User-dependent composition;
Occurrence number in the composition related to described targeted customer is more than the composition of the 3rd preset value as described target
The corresponding composition of user;
Described targeted customer corresponding first composition list is generated according to the corresponding composition of described targeted customer.
For said method, in a kind of possible implementation, the user behavior data of collection targeted customer, according to institute
State user behavior data and generate the first multimedia resource list, including:
All user behavior datas of the described targeted customer in the time period are specified in collection;
Effective user behavior data is filtered out from the user behavior data being gathered;
According to the described effective user behavior data corresponding time, described effective user behavior data is ranked up,
Obtain described first multimedia resource list.
In order to solve above-mentioned technical problem, according to another embodiment of the present invention, there is provided a kind of determination multimedia resource
Title device, including:
Acquisition module, for gathering the user behavior data of targeted customer, and generates the according to described user behavior data
One multimedia resource list;
First parsing module, for solving to the title of each multimedia resource in described first multimedia resource list
Analysis, obtains described targeted customer corresponding first composition list;
Second parsing module, for parsing to the former title of multimedia resource to be recommended, obtains described former title pair
The second composition list answered;
Comparison module, for by each in each composition in described second composition list and described first composition list
Composition is compared, the second composition list after being updated;
Determining module, for determining the new of described multimedia resource to be recommended according to the second composition list after described renewal
Title.
For said apparatus, in a kind of possible implementation, described comparison module includes:
Similarity Measure submodule, for calculating each composition in described second composition list and described first composition row
The similarity of each composition in table;
Replace submodule, for becoming with described first composition list in the composition in described second composition list
In the case that the similarity divided is more than the first preset value, the composition in described second composition list is replaced with described the first one-tenth
Divide the composition in list;
Update submodule, for the second composition list after being updated according to the composition of all replacements.
For said apparatus, in a kind of possible implementation, described Similarity Measure submodule includes:
Vector determination unit, for determining the corresponding vector of each composition in described second composition list;
Similarity calculated, for calculating the corresponding vector of each composition in described second composition list and institute respectively
State the similarity of the corresponding vector of each composition in first composition list.
For said apparatus, in a kind of possible implementation, described similarity calculated is used for:
Calculate the corresponding vector of l-th composition in described second composition list using formula 1With described first composition list
In the corresponding vector of m-th compositionSimilarity
For said apparatus, in a kind of possible implementation, described determining module includes:
Score calculating sub module, for calculating the score of the second composition list after described renewal;
Determination sub-module, for described update after second composition list score be more than the second preset value situation
Under, the new title of described multimedia resource to be recommended is determined according to the second composition list after described renewal.
For said apparatus, in a kind of possible implementation, described score calculating sub module is used for:
The probability meter being occurred in specified sample set according to each composition in the second composition list after described renewal
Calculate the score of the second composition list after described renewal.
For said apparatus, in a kind of possible implementation, described score calculating sub module is used for:
Calculate score s of the second composition list after described renewal using formula 2;
Wherein, the number of composition, w in the second composition list after n represents described renewaljRepresent second after described renewal
J-th composition in ingredient lists, wj-iRepresent jth-i composition in the second composition list after described renewal, p (wjwj-i)
Represent described j-th composition and described jth-i the composition common probability occurring, p (w in described specified sample setj-i) table
Show the probability that described jth-i composition occurs in described specified sample set.
For said apparatus, in a kind of possible implementation, described device also includes:
Reservation module, the score for the second composition list after described renewal is less than or equal to described second preset value
In the case of, retain the former title of described multimedia resource to be recommended.
For said apparatus, in a kind of possible implementation, described first parsing module includes:
Analyzing sub-module, for solving to the title of each multimedia resource in described first multimedia resource list
Analysis, obtains the composition related to described targeted customer;
Composition determination sub-module, for being more than the 3rd preset value by occurrence number in the composition related to described targeted customer
Composition as the corresponding composition of described targeted customer;
First composition list generates submodule, for generating described targeted customer according to the corresponding composition of described targeted customer
Corresponding first composition list.
For said apparatus, in a kind of possible implementation, described acquisition module includes:
Collection submodule, for gathering all user behavior datas of the described targeted customer in the specified time period;
Screening submodule, for filtering out effective user behavior data from the user behavior data being gathered;
Sorting sub-module, for according to the described effective user behavior data corresponding time to described effective user's row
It is ranked up for data, obtain described first multimedia resource list.
Beneficial effect
Used with target by each composition in the former title corresponding second composition list by multimedia resource to be recommended
Each composition in the corresponding first composition list of family is compared, and the second composition list after being updated, so that it is determined that treat
Recommend the new title of multimedia resource, the method and device of the title of determination multimedia resource according to embodiments of the present invention can
Determine personalized title for targeted customer, preferably can attract user such that it is able to improve recommended multimedia money
The clicked probability in source.
According to below with reference to the accompanying drawings, to detailed description of illustrative embodiments, the further feature of the present invention and aspect will become
Clear.
Brief description
Comprise in the description and constitute the accompanying drawing of a part of specification and specification together illustrates the present invention's
Exemplary embodiment, feature and aspect, and for explaining the principle of the present invention.
Fig. 1 illustrates the flowchart of the method for title determining multimedia resource according to an embodiment of the invention;
Fig. 2 illustrates an example of the method and step S104 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property;
Fig. 3 illustrates an example of the method and step S301 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property;
Fig. 4 illustrates an example of the method and step S105 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property;
Fig. 5 illustrates an example of the method and step S102 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property;
Fig. 6 illustrates an example of the method and step S101 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property;
Fig. 7 illustrates the structured flowchart of the device of title determining multimedia resource according to another embodiment of the present invention;
Fig. 8 illustrates the one of the structured flowchart of the device of title determining multimedia resource according to another embodiment of the present invention
Exemplary flowchart;
Fig. 9 shows a kind of structural frames of the equipment of title of determination multimedia resource of an alternative embodiment of the invention
Figure.
Specific embodiment
Describe various exemplary embodiments, feature and the aspect of the present invention below with reference to accompanying drawing in detail.Identical in accompanying drawing
Reference represent the same or analogous element of function.Although the various aspects of embodiment shown in the drawings, remove
Non-specifically points out it is not necessary to accompanying drawing drawn to scale.
Special word " exemplary " means " as example, embodiment or illustrative " here.Here as " exemplary "
Illustrated any embodiment should not necessarily be construed as preferred or advantageous over other embodiments.
In addition, in order to better illustrate the present invention, giving numerous details in specific embodiment below.
It will be appreciated by those skilled in the art that not having some details, the present invention equally can be implemented.In some instances, for
Method well known to those skilled in the art, means, element and circuit are not described in detail, in order to highlight the purport of the present invention.
Embodiment 1
Fig. 1 illustrates the flowchart of the method for title determining multimedia resource according to an embodiment of the invention.Should
The executive agent of embodiment can for server or other determine multimedia resource title device, here do not make
Limit.As shown in figure 1, the method mainly includes:
In step S101, the user behavior data of collection targeted customer, and generate more than first according to user behavior data
Media Resource List.
Wherein, multimedia can be the synthesis of media, for example, can include the medias such as word, sound and image
Form.For example, multimedia resource can be video, is not limited thereto.The user behavior data of targeted customer can include but
Be not limited to following at least one:Targeted customer watches the data of multimedia resource, targeted customer comment on multimedia resource data,
Targeted customer subscribes to the data of multimedia resource and the data of multimedia resource is stepped on targeted customer top.In the present embodiment, permissible
The corresponding multimedia resource of user behavior data according to targeted customer generates the first multimedia resource list.For example, target is used
The corresponding first multimedia resource list in family can be expressed as LU={ v1, v2 ..., vn }.
In step s 102, the title of each multimedia resource in the first multimedia resource list is parsed, obtain
Targeted customer's corresponding first composition list.
As an example of the present embodiment, can be using NER (Named Entity Recognition, name entity
Identification) technology parses to the title of each multimedia resource in the first multimedia resource list, to obtain targeted customer couple
The first composition list answered.
In step s 103, the former title of multimedia resource to be recommended is parsed, obtain former title corresponding second
Ingredient lists.
As an example of the present embodiment, can be to be recommended many to each in multimedia resource list to be recommended respectively
The former title of media resource is parsed, and obtains the corresponding second composition list of each former title.It is for instance possible to use NER skill
Art parses to the former title of multimedia resource to be recommended, obtains the corresponding second composition list of former title.
In step S104, each composition in second composition list is carried out with each composition in first composition list
Relatively, the second composition list after being updated.
As an example of the present embodiment, can respectively each composition in second composition list be arranged with first composition
Each composition in table is compared, to replace the composition in second composition list using the composition in first composition list.
The new title of multimedia resource to be recommended in step S105, is determined according to the second composition list after updating.
For example, former entitled " tortoise gnaws the toe of the Pussy of a sleep " of multimedia resource to be recommended, new mark
It is entitled that " tortoise gnaws the toe of the mew star people of a sleep!”.
The present embodiment passes through each composition in the former title corresponding second composition list of multimedia resource to be recommended
Each composition in first composition list corresponding with targeted customer is compared, the second composition list after being updated, from
And determine the new title of multimedia resource to be recommended, personalized title can be determined for targeted customer, can preferably inhale
Quote family such that it is able to improve the clicked probability of recommended multimedia resource;Mark without manual amendment's multimedia resource
Topic, greatlys save human cost.
Fig. 2 illustrates an example of the method and step S104 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property.As shown in Fig. 2 by each composition in each composition in second composition list and first composition list
It is compared, the second composition list after being updated, including:
In step s 201, calculate each composition in each composition and the first composition list in second composition list
Similarity.
For example, it is possible to the similarity between composition is determined by the similarity between the corresponding vector of calculating composition.This
Skilled person should be understood that and can also weigh similarity between composition by other parameters of composition, here does not limit
Fixed.
In step S202, the similarity of the composition in the composition in second composition list with first composition list
In the case of the first preset value, the composition in second composition list is replaced with the composition in first composition list.
For example, the first preset value can be 0.9.For example, the composition in second composition list is " Pussy ", the first one-tenth
The composition in list is divided to be " mew star people ", " Pussy " is 0.95 with the similarity of " mew star people ", then can be by second composition
" Pussy " in list replaces with " mew star people " in first composition list.
In this example, the similarity of the composition in the composition in second composition list with first composition list is big
In the case of the first preset value, just the composition in second composition list is replaced with the composition in first composition list,
Thus, it is possible to ensure the uniformity of semanteme.
Second composition list in step S203, after being updated according to the composition of all replacements.
Fig. 3 illustrates an example of the method and step S201 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property.As shown in figure 3, each composition calculating in second composition list is become with each in first composition list
The similarity divided, including:
In step S301, determine the corresponding vector of each composition in second composition list.
As an example of the present embodiment, each composition pair in second composition list can be determined using word2vec
The corresponding vector of each composition in the vector answered and first composition list.
In step s 302, the corresponding vector of each composition in second composition list and first composition list are calculated respectively
In each composition corresponding vector similarity.
For example, it is possible to the COS distance between corresponding for two compositions vector is defined as the similarity of two compositions.
In a kind of possible implementation, calculate the corresponding vector of each composition and the in second composition list respectively
The similarity of the corresponding vector of each composition in one ingredient lists, including:L in second composition list is calculated using formula 1
The corresponding vector of individual compositionWith the corresponding vector of m-th composition in first composition listSimilarity
Fig. 4 illustrates an example of the method and step S105 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property.As shown in figure 4, determine the new mark of multimedia resource to be recommended according to the second composition list after updating
Topic, including:
In step S401, calculate the score of the second composition list after updating.
In step S402, in the case that the score of second composition list in the updated is more than the second preset value, according to
Second composition list after renewal determines the new title of multimedia resource to be recommended.
In this example, in the case that the score of second composition list in the updated is more than the second preset value, according to more
Second composition list after new determines the new title of multimedia resource to be recommended, with ensure new title before and after language between composition
Speech relevance.Wherein, the second preset value can set according to the experience of those skilled in the art, is not limited thereto.
In a kind of possible implementation, after the score of the second composition list calculating after updating, the method is also
Including:In the case that the score of second composition list in the updated is less than or equal to the second preset value, retain many matchmakers to be recommended
The former title of body resource.In this implementation, the score of second composition list in the updated is less than or equal to second and presets
In the case of value, retain the former title of multimedia resource to be recommended, with ensure title before and after language relevance between composition.
In a kind of possible implementation, calculate the score of the second composition list after updating, including:After updating
Second composition list in the probability calculation that occurs in specified sample set of each composition update after second composition list
Score.
Specify sample for example, it is possible to determine according to the title of all multimedia resources in multimedia resource list to be recommended
Set, or specified sample set can be determined according to the title of all multimedia resources in the multimedia resource list that other are specified
Close, be not limited thereto.
In a kind of possible implementation, according to each composition in the second composition list after updating in specified sample
The score of the second composition list after the probability calculation renewal occurring in set, including:
Calculate score s of the second composition list after updating using formula 2;
Wherein, n represents the number of composition in the second composition list after renewal, wjRepresent the second composition list after updating
In j-th composition, wj-iRepresent jth-i composition in the second composition list after updating, p (wjwj-i) represent j-th one-tenth
Divide the probability with the common appearance in specified sample set of jth-i composition, p (wj-i) represent jth-i composition in specified sample
The probability occurring in set.
Fig. 5 illustrates an example of the method and step S102 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property.As shown in figure 5, parsing to the title of each multimedia resource in the first multimedia resource list,
Obtain targeted customer's corresponding first composition list, including:
In step S501, the title of each multimedia resource in the first multimedia resource list is parsed, obtains
The composition related to targeted customer.
As an example of the present embodiment, can using NER technology respectively in the first multimedia resource list each
The title of multimedia resource is parsed, and obtains the corresponding composition of title of each multimedia resource.Wherein, composition can include
Entity word (such as " dog " " Mars intelligence bureau "), mood word (such as " good-looking " " terribly funny do not pay with one's life ") and mood punctuate is (for example
“!") in one or more.Wherein, entity word can include one or many in name, place name, mechanism's name and proper noun
?.
In step S502, using occurrence number in the composition related to targeted customer be more than the 3rd preset value composition as
The corresponding composition of targeted customer.
For example, the 3rd preset value can be 2.In this example, by arranging the 3rd preset value, will be related to targeted customer
Composition in occurrence number be more than the composition of the 3rd preset value as the corresponding composition of targeted customer, and filter and targeted customer's phase
In the composition closing, occurrence number is less than or equal to the composition of the 3rd preset value, corresponding to targeted customer thus, it is possible to reduce noise
The impact of composition.
In step S503, targeted customer's corresponding first composition list is generated according to the corresponding composition of targeted customer.
For example, the list of targeted customer's corresponding first composition can be expressed as { NE1, NE2 ..., NEn }, wherein, NE1,
NE2 ..., NEn represents targeted customer's each composition corresponding.
Fig. 6 illustrates an example of the method and step S101 of title determining multimedia resource according to an embodiment of the invention
The flowchart of property.As shown in fig. 6, the user behavior data of collection targeted customer, generate first according to user behavior data
Multimedia resource list, including:
In step s 601, all user behavior datas specifying the targeted customer in the time period are gathered.
For example, it is possible to gather all user behavior datas of the targeted customer in 1 month, 3 months or half a year.
In step S602, filter out effective user behavior data from the user behavior data being gathered.
For example, it is possible to the user behavior data repeating to watch multimedia resource is defined as invalid user behavior data,
The user behavior data of the completed percentage very little of viewing multimedia resource can also be defined as invalid user behavior data,
This is not construed as limiting.
In step S603, according to the effective user behavior data corresponding time, effective user behavior data is carried out
Sequence, obtains the first multimedia resource list.
Wherein, when the effective user behavior data corresponding time can be the generation of this effective user behavior data
Between.According to the effective user behavior data corresponding time, effective user behavior data is ranked up can be:According to having
The user behavior data of effect time sequencing from the near to the remote is ranked up to effective user behavior data.
In a kind of possible implementation, multimedia resource list to be recommended can be screened, so as to be recommended
Multimedia resource possesses diversity:Channel belonging to the uploader information of multimedia resource to be recommended, multimedia resource to be recommended
Information, targeted customer's viewing data of multimedia resource and the interest tags of targeted customer.For example, if multimedia resource to be recommended
List includes the multimedia resource that more than four same uploader upload, then can retain the multimedia money of this uploader upload
In source, the click volume ranking multimedia resource of first three is as multimedia resource to be recommended.Again for example, if multimedia resource to be recommended
List includes the multimedia resource of more than four same two grades of channels, then can retain in the multimedia resource of this two grades of channels
The click volume ranking multimedia resource of first three is as multimedia resource to be recommended.For example, XATV-6 is a certain level channel, lake
Southern XATV-6 is two grades of channels under this level channel.Again for example, if multimedia resource list to be recommended include four with
Multimedia resource under upper same three-level interest tags, then can retain in the multimedia resource under this three-level interest tags and click on
The amount ranking multimedia resource of first three is as multimedia resource to be recommended.For example, one-level interest tags are amusement, and star in amusement circle is
Two grades of interest tags under this one-level interest tags, Beyond is the three-level interest tags under this two grades of interest tags.Again for example,
If multimedia resource list to be recommended includes the multimedia resource that targeted customer watched in the recent period, not by this multimedia resource
As multimedia resource to be recommended.
So, by each composition in the former title corresponding second composition list by multimedia resource to be recommended and mesh
Each composition in mark user's corresponding first composition list is compared, the second composition list after being updated, thus really
The new title of fixed multimedia resource to be recommended, the method for the title of determination multimedia resource according to embodiments of the present invention being capable of pin
Personalized title is determined to targeted customer, preferably can attract user such that it is able to improve recommended multimedia resource
Clicked probability.
Embodiment 2
Fig. 7 illustrates the structured flowchart of the device of title determining multimedia resource according to another embodiment of the present invention.Fig. 7
The method that shown device can be used for running the title of determination multimedia resource shown in Fig. 1 to Fig. 6.For convenience of description,
Illustrate only part related to the present embodiment in the figure 7.
As shown in fig. 7, this device includes:Acquisition module 71, for gathering the user behavior data of targeted customer, and according to
Described user behavior data generates the first multimedia resource list;First parsing module 72, for described first multimedia money
In the list of source, the title of each multimedia resource is parsed, and obtains described targeted customer corresponding first composition list;Second
Parsing module 73, for parsing to the former title of multimedia resource to be recommended, obtains corresponding the second one-tenth of described former title
Divide list;Comparison module 74, for will be each in each composition in described second composition list and described first composition list
Individual composition is compared, the second composition list after being updated;Determining module 75, for according to the second one-tenth after described renewal
List is divided to determine the new title of described multimedia resource to be recommended.
Fig. 8 illustrates the one of the structured flowchart of the device of title determining multimedia resource according to another embodiment of the present invention
Exemplary flowchart.Device shown in Fig. 8 can be used for running the mark of the determination multimedia resource shown in Fig. 1 to Fig. 6
The method of topic.For convenience of description, illustrate only part related to the present embodiment in fig. 8.In Fig. 8, label is identical with Fig. 7
Assembly there is identical function, for simplicity's sake, omit detailed description to these assemblies.
In a kind of possible implementation, described comparison module 74 includes:Similarity Measure submodule 741, by based on
Calculate the similarity of each composition in described second composition list and each composition in described first composition list;Replace submodule
Block 742, the similarity for the composition in the composition in described second composition list with described first composition list is big
In the case of the first preset value, the composition in described second composition list is replaced with described first composition list
Composition;Update submodule 743, for the second composition list after being updated according to the composition of all replacements.
In a kind of possible implementation, described Similarity Measure submodule 741 includes:Vector determination unit, is used for
Determine the corresponding vector of each composition in described second composition list;Similarity calculated, for calculating described respectively
Each composition in binary list is corresponding vectorial and described first composition list in the corresponding vector of each composition phase
Like degree.
In a kind of possible implementation, described similarity calculated is used for:Described second composition is calculated using formula 1
The corresponding vector of l-th composition in listWith the corresponding vector of m-th composition in described first composition listSimilar
Degree
In a kind of possible implementation, described determining module 75 includes:Score calculating sub module 751, for calculating
The score of the second composition list after described renewal;Determination sub-module 752, for the second composition list after described renewal
In the case that score is more than the second preset value, described multimedia money to be recommended is determined according to the second composition list after described renewal
The new title in source.
In a kind of possible implementation, described score calculating sub module 751 is used for:According to second after described renewal
Second composition list after updating described in the probability calculation that each composition in ingredient lists occurs in specified sample set
Score.
In a kind of possible implementation, described score calculating sub module 751 is used for:Described renewal is calculated using formula 2
Score s of second composition list afterwards;
Wherein, the number of composition, w in the second composition list after n represents described renewaljRepresent second after described renewal
J-th composition in ingredient lists, wj-iRepresent jth-i composition in the second composition list after described renewal, p (wjwj-i)
Represent described j-th composition and described jth-i the composition common probability occurring, p (w in described specified sample setj-i) table
Show the probability that described jth-i composition occurs in described specified sample set.
In a kind of possible implementation, described device also includes:Reservation module 76, for the after described renewal
In the case that the score of binary list is less than or equal to described second preset value, retain the former of described multimedia resource to be recommended
Title.
In a kind of possible implementation, described first parsing module 72 includes:Analyzing sub-module 721, for institute
The title stating each multimedia resource in the first multimedia resource list is parsed, and obtains the one-tenth related to described targeted customer
Point;Composition determination sub-module 722, for being more than the 3rd preset value by occurrence number in the composition related to described targeted customer
Composition is as the corresponding composition of described targeted customer;First composition list generates submodule 723, for according to described targeted customer
Corresponding composition generates described targeted customer corresponding first composition list.
In a kind of possible implementation, described acquisition module 71 includes:Collection submodule 711, specified for gathering
All user behavior datas of the described targeted customer in the time period;Screening submodule 712, for from the user behavior being gathered
Effective user behavior data is filtered out in data;Sorting sub-module 713, for according to described effective user behavior data pair
The time answered is ranked up to described effective user behavior data, obtains described first multimedia resource list.
It should be noted that so, by the former title corresponding second composition list by multimedia resource to be recommended
The first composition list corresponding with targeted customer of each composition in each composition be compared, the second one-tenth after being updated
Point list, so that it is determined that the new title of multimedia resource to be recommended, the mark of determination multimedia resource according to embodiments of the present invention
The device of topic can determine personalized title for targeted customer, preferably can attract user and be pushed away such that it is able to improve
The clicked probability of the multimedia resource recommended.
Embodiment 3
Fig. 9 shows a kind of structural frames of the equipment of title of determination multimedia resource of an alternative embodiment of the invention
Figure.The equipment 1100 of the described title determining multimedia resource can be host server, the individual calculus possessing computing capability
Machine PC or portable portable computer or terminal etc..The specific embodiment of the invention not concrete reality to calculate node
Now limit.
The equipment 1100 of the described title determining multimedia resource includes processor (processor) 1110, communication interface
(Communications Interface) 1120, memory (memory) 1130 and bus 1140.Wherein, processor 1110,
Communication interface 1120 and memory 1130 complete mutual communication by bus 1140.
Communication interface 1120 is used for and network device communications, and wherein the network equipment includes such as Virtual Machine Manager center, is total to
Enjoy storage etc..
Processor 1110 is used for configuration processor.Processor 1110 is probably a central processor CPU, or special collection
Become circuit ASIC (Application Specific Integrated Circuit), or be arranged to implement the present invention
One or more integrated circuits of embodiment.
Memory 1130 is used for depositing file.Memory 1130 may comprise high-speed RAM memory it is also possible to also include non-
Volatile memory (non-volatile memory), for example, at least one magnetic disc store.Memory 1130 can also be deposited
Memory array.Memory 1130 is also possible to by piecemeal, and described piece can be combined into virtual volume by certain rule.
In a kind of possible embodiment, said procedure can be the program code including computer-managed instruction.This journey
Sequence is particularly used in:Realize the operation of each step in embodiment 1.
Those of ordinary skill in the art are it is to be appreciated that each exemplary cell in embodiment described herein and algorithm
Step, being capable of being implemented in combination in electronic hardware or computer software and electronic hardware.These functions are actually with hardware also
Being software form to realize, the application-specific depending on technical scheme and design constraint.Professional and technical personnel can be directed to
Specifically application selects different methods to realize described function, but this realization is it is not considered that exceed the model of the present invention
Enclose.
If to be realized using in the form of computer software described function and as independent production marketing or use when,
To a certain extent it is believed that all or part (part for example prior art being contributed) of technical scheme is
Embody in form of a computer software product.This computer software product is generally stored inside the non-volatile of embodied on computer readable
In storage medium, including some instructions with so that computer equipment (can be that personal computer, server or network set
Standby etc.) all or part of step of execution various embodiments of the present invention method.And aforesaid storage medium include USB flash disk, portable hard drive,
Read-only storage (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic
Dish or CD etc. are various can be with the medium of store program codes.
The above, the only specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, and any
Those familiar with the art the invention discloses technical scope in, change or replacement can be readily occurred in, all should contain
Cover within protection scope of the present invention.Therefore, protection scope of the present invention should be defined by described scope of the claims.