CN106776697A - Content recommendation method and device - Google Patents
Content recommendation method and device Download PDFInfo
- Publication number
- CN106776697A CN106776697A CN201610997930.8A CN201610997930A CN106776697A CN 106776697 A CN106776697 A CN 106776697A CN 201610997930 A CN201610997930 A CN 201610997930A CN 106776697 A CN106776697 A CN 106776697A
- Authority
- CN
- China
- Prior art keywords
- content
- interest
- targeted customer
- alternating
- degree
- 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.)
- Pending
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
- G06F16/90—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)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the present application provides a kind of content recommendation method and device.Content recommendation method includes：From the multiple network behavioral data of targeted customer, the interest content of targeted customer is extracted；According to interest content, interest-degree of the analysis targeted customer to alternating content；According to interest-degree, the selection target content from alternating content；Recommend object content to targeted customer.The method provided using the embodiment of the present application so that the content for a purpose recommended is stronger, and the probability for meeting user's request is higher, can improve user to recommending the Experience Degree of behavior, be conducive to exciting user interest, reach recommendation purpose.
Description
Technical field
The application is related to Internet technical field, more particularly to a kind of content recommendation method and device.
Background technology
With the development of Internet technology, the application based on internet is more and more.In order to improve the popularity of application with
And preferably service user, a critical function of application is had become to user's content recommendation.
The existing way of recommendation is more according to certain standard, such as user's sex, the hobby label of setting or operation
Demand etc., to user's content recommendation.
The content of the invention
During commending contents are done according to the existing way of recommendation, inventor has found：The effect of existing commending contents mode
It is really less desirable, it is impossible to produce a desired effect.
Studied by initial analysis, inventor has found that the major reason for causing recommendation effect undesirable is：That is recommended is interior
Hold specific aim poor, be not user real interested, it is impossible to excite the point of interest of user.Because not being interested interior of user
Hold, so user will not pay close attention to or can directly delete recommended content, what is more, can directly close or shielding push away
Recommend function.
By further analysis and research, inventor find it is existing apply do recommend when more attention be recommended requirements and
Purpose, and seldom pay close attention to user's request and impression, so causing the content recommendation cannot to excite the point of interest of user, it is impossible to reach and push away
Recommend purpose.
If commending contents can be carried out with reference to user interest and impression while recommended requirements are met, this can be very big
Improve recommendation effect in ground.But how to consider that user interest and impression turn into the problem for needing to solve in recommendation process.
Found by studying inventor：The network behavior that user has occurred to a certain extent, embodies the interest of user
Preference, therefore the network behavior data of user can be based on, determine user's content interested；And then, it is interested based on user
Content, goes to analyze interest-degree of the user to other contents；Based on user to the interest-degree of other contents, to user's content recommendation.
Interest of the way of recommendation from user so that the content for a purpose recommended is stronger, the probability for meeting user's request is higher,
User can be improved to recommending the Experience Degree of behavior, be conducive to exciting user interest, reach recommendation purpose.
Based on above-mentioned analysis, the embodiment of the present application provides a kind of content recommendation method, including：
From the multiple network behavioral data of targeted customer, the interest content of targeted customer is extracted；
According to the interest content, interest-degree of the targeted customer to alternating content is analyzed；
According to the interest-degree, the selection target content from the alternating content；
Recommend the object content to the targeted customer.
Optionally, the analytical procedure of the interest-degree, including：
Analyze the feature association degree of the alternating content and the interest content；
According to the feature association degree, interest-degree of the targeted customer to the alternating content is determined.
Optionally, the analytical procedure of the feature association degree, including：
From the interest content, the interest characteristics set of the targeted customer is extracted；
For each interest characteristics in the interest characteristics set configures weight；
Analyze the interest characteristics associated with the alternating content in the interest characteristics set；
According to the weight of the interest characteristics associated with the alternating content, the feature association degree is obtained.
Optionally, the configuration step of the weight, including：
It is the interest characteristics configuration weight according to the click volume of the interest characteristics.
Optionally, the analytical procedure of the interest characteristics associated with the alternating content, including：
The feature in the alternating content is extracted, as the candidate feature；
The candidate feature is matched in the interest characteristics set；
Obtain the interest characteristics in the candidate feature matching.
Optionally, the selection step of the object content, including：
According to the interest-degree, from the alternating content, the diversification ground selection object content.
Optionally, the recommendation step of the object content, including：
With polymerization methodses, the object content of the Multiple strategies is recommended to the targeted customer.
Optionally, before the interest-degree is analyzed, also include：
Directly by the content in platform database, as the alternating content；And/or
Similar users according to targeted customer content interested, determines the alternating content.
Optionally, methods described also includes：
Similarity in analysis user's set between each user content interested and the interest content；
According to the similarity, the similar users of the targeted customer are obtained from user set.
Optionally, the determination step of the alternating content, including：
The feature in similar users content interested is extracted, as similar features；
From the interest content, the interest characteristics set of the targeted customer is extracted；
Obtain the distinguishing characteristics that the interest characteristics set is not belonging in the similar features；
From the network behavior data or platform database of the similar users, obtain interior with the distinguishing characteristics
Hold, as the alternating content.
Correspondingly, the embodiment of the present application also provides a kind of content recommendation device, including：
Extraction module, for from the multiple network behavioral data of targeted customer, extracting in the interest of the targeted customer
Hold；
Analysis module, for according to the interest content, analyzing interest-degree of the targeted customer to alternating content；
Selecting module, for according to the interest-degree, the selection target content from the alternating content；
Recommending module, for recommending the object content to the targeted customer.
Optionally, the analysis module includes：
Analytic unit, the feature association degree for analyzing the alternating content and the interest content；
Determining unit, for according to the feature association degree, determining interest of the targeted customer to the alternating content
Degree.
Optionally, the analytic unit specifically for：
From the interest content, the interest characteristics set of the targeted customer is extracted；
For each interest characteristics in the interest characteristics set configures weight；
Analyze the interest characteristics associated with the alternating content in the interest characteristics set；
According to the weight of the interest characteristics associated with the alternating content, the feature association degree is obtained.
Optionally, the selecting module specifically for：
According to the interest-degree, from the alternating content, the diversification ground selection object content.
Optionally, the recommending module specifically for：
With polymerization methodses, the object content of the Multiple strategies is recommended to the targeted customer.
Optionally, described device also includes：
Determining module, for the content in direct access platform database, as the alternating content；And/or, according to institute
The similar users of targeted customer content interested is stated, the alternating content is determined.
Optionally, the determining module specifically for：
The feature in similar users content interested is extracted, as similar features；
From the interest content, the interest characteristics set of the targeted customer is extracted；
Obtain the distinguishing characteristics that the interest characteristics set is not belonging in the similar features；
From the network behavior data or platform database of the similar users, obtain interior with the distinguishing characteristics
Hold, as the alternating content.
In the embodiment of the present application, the multiple network behavioral data based on user, determines user's content interested；Enter
And, based on user's content interested, go to analyze interest-degree of the user to alternating content；Based on user to the interest of alternating content
Degree, to user's content recommendation.Interest of this way of recommendation from user so that the content for a purpose recommended is stronger, symbol
The probability for closing user's request is higher, can improve user to recommending the Experience Degree of behavior, is conducive to exciting user interest, reaches and pushes away
Recommend purpose.
Brief description of the drawings
Accompanying drawing described herein is used for providing further understanding of the present application, constitutes the part of the application, this Shen
Schematic description and description please does not constitute the improper restriction to the application for explaining the application.In the accompanying drawings：
The schematic flow sheet of the content recommendation method that Fig. 1 is provided for the embodiment of the application one；
The schematic flow sheet of the content recommendation method that Fig. 2 is provided for another embodiment of the application；
The schematic flow sheet of the content recommendation method that Fig. 3 is provided for the another embodiment of the application；
The structural representation of the content recommendation device that Fig. 4 is provided for the another embodiment of the application；
The schematic flow sheet of the content recommendation device that Fig. 5 is provided for the another embodiment of the application.
Specific embodiment
To make the purpose, technical scheme and advantage of the application clearer, below in conjunction with the application specific embodiment and
Corresponding accompanying drawing is clearly and completely described to technical scheme.Obviously, described embodiment is only the application one
Section Example, rather than whole embodiments.Based on the embodiment in the application, those of ordinary skill in the art are not doing
Go out the every other embodiment obtained under the premise of creative work, belong to the scope of the application protection.
The schematic flow sheet of the content recommendation method that Fig. 1 is provided for the embodiment of the application one.As shown in figure 1, the method bag
Include：
101st, from the multiple network behavioral data of targeted customer, the interest content of targeted customer is extracted.
102nd, according to the interest content, interest-degree of the analysis targeted customer to alternating content.
103rd, according to the interest-degree, the selection target content from alternating content.
104th, the object content is recommended to targeted customer.
In Internet era, user can produce multiple network behavior by its terminal, for example, browse, buy, paying close attention to, adding purchase
The network behaviors such as thing car, subscription.By research, inventor has found：Data (the referred to as network behavior that these network behaviors are produced
Data) interest preference of user can be embodied to a certain extent.For example, user often browses or buy a certain commodity, illustrate to use
Family may be interested be in the commodity；Again for example, user subscribes to or often browses life leisure article, illustrate user to this
A little articles are interested；Again for example, user pays close attention to certain shop, illustrate interested in the commodity of the shop operation, etc..
Based on above-mentioned discovery, the present embodiment provides a kind of content recommendation method.For ease of describing and distinguishing, the present embodiment with
As a example by targeted customer, the flow of the content recommendation method that the present embodiment is provided is illustrated.The targeted customer can be any user.
When needing to targeted customer's content recommendation, the multiple network behavioral data related to targeted customer can be obtained,
The acquired network behavior data of analysis, therefrom extract targeted customer's content (referred to as interest content) interested.
In the present embodiment, interest content not directly according to targeted customer, similar content is recommended to targeted customer,
Because Similar content of the targeted customer not necessarily to interest content is interested, if directly the Similar content of interest content recommended
To targeted customer, it is possible to interfered to targeted customer, using experience degree of the targeted customer to application is reduced, it is impossible to reach and push away
Recommend purpose.
In the present embodiment, the interest content according to targeted customer, analyzes interest-degree of the targeted customer to alternating content, then
According to interest-degree, the selection target content from alternating content recommends selected object content to targeted customer.The present embodiment from
The interest of targeted customer is set out so that the content for a purpose recommended is stronger, and the probability for meeting user's request is higher, is conducive to drop
Low recommendation behavior brings the probability of interference to user, can improve user to recommending the Experience Degree of behavior, is conducive to exciting user
Interest, reaches recommendation purpose.
In above-described embodiment or following embodiments, using multiple network behavioral data, can carry more comprehensively, exactly
Take the interest content of targeted customer.The species of the network behavior data is not limited to：Subscribe to, buy, browsing, paying close attention to, adding shopping
Car, payment, download etc..Wherein, every kind of network behavior data are included but is not limited to：Network behavior referent (such as business
Product, article, comment, film, music, shop etc.), the corresponding user profile of network behavior, network behavior occur time, network
Information of terminal device etc. where place, the property of network behavior, network behavior that behavior occurs.
Optionally, it is possible to use the subnetwork behavioral data that targeted customer produces, can for example be existed using targeted customer
The various network behavior data produced in the time period are specified, for example, the various networks produced in nearest month or a week
Behavioral data.Or, it is possible to use all network behavior data for getting of targeted customer.
Optionally, every kind of network behavior data can be individually analyzed, the interest content of targeted customer is therefrom extracted.Example
Such as, analyze the buying behavior of targeted customer, it is assumed that targeted customer have purchased the household electrical appliances such as TV, refrigerator, washing machine in the recent period, then can be with
The interest content for extracting targeted customer is household electrical appliances.For example, the subscription behavior of analysis targeted customer, it is assumed that targeted customer have subscribed food
The health-related article of product, the interest content that can extract targeted customer is health care of food class article.For example, analysis targeted customer's is clear
Look at behavior, it is assumed that targeted customer frequently browses menu in the recent period, the interest content that can extract targeted customer is menu.This mode
Can more extensively, comprehensively extract the interest content of targeted customer.Or
Optionally, multiple network behavioral data can be associatedly analyzed, the interest content of targeted customer is therefrom extracted.Example
Such as, the associatedly navigation patterns of analysis targeted customer, buying behavior and concern behavior, it is assumed that targeted customer frequently browses in the recent period
Boot, and cotton boots are have purchased in the recent period, while having paid close attention to the brand shop of sale boots, can extract in the interest of targeted customer
Hold is boot or cotton boots.For example, the associatedly navigation patterns of analysis targeted customer, buying behavior and subscription behavior, it is assumed that
Targeted customer frequently browses menu in the recent period, and have purchased in the recent period fresh, while have subscribed the articles such as cuisines story, can extract
The interest content of targeted customer is cuisines or menu etc..This mode can improve the accuracy of extracted interest content.
In above-described embodiment or following embodiments, according to interest content, targeted customer is to the emerging of alternating content for analysis
During interesting degree, concrete application scene can be combined.According to different application scene, the analysis mode of interest-degree may not
Together.
Optionally, in application scenes, the content that targeted customer may be similar to content details is interested.Cause
This, can be based on interest of the targeted customer to interest content, analysis alternating content and similarity of the interest content in content details,
The similarity in content details is higher with interest content for alternating content, it is meant that targeted customer to the interest-degree of alternating content also
It is higher.Then according to interest content, analysis targeted customer can be to the interest-degree of alternating content：In analysis alternating content and interest
The details similarity of appearance, according to details similarity, determines interest-degree of the targeted customer to alternating content.
Optionally, in application scenes, targeted customer may more pay close attention to relevance of two contents in certain feature.
For example, it is assumed that targeted customer have purchased it is fresh, may be pair interested to fresh related menu or article.And for example, it is assumed that
Targeted customer has browsed health care of food class encyclopaedia, and the relevant healthcare product that may be introduced the health care of food class encyclopaedia are interested.
Therefore, interest of the targeted customer to interest content, analysis alternating content and the interest content degree of association characteristically can be based on, be waited
Select the content degree of association with interest content characteristically higher, it is meant that targeted customer also gets over to the interest-degree of alternating content
It is high.Then according to interest content, analysis targeted customer can be to the interest-degree of alternating content：Analysis alternating content and interest content
Feature association degree；According to feature association degree, interest-degree of the targeted customer to alternating content is determined.
What deserves to be explained is, when analyzing feature association and spending, do not limit characteristic type, visual concrete application scene and should
Depending on demand.
For example, the degree of association of alternating content and interest content on functional character can be analyzed.The functional character is main
Refer to what function is content have, or can be used for what does.For example, fresh can be used to do cuisines, recipe can be used to instruct to do
Cuisines, both have certain association on functional character.Based on this, if targeted customer have purchased fresh, can be used to target
Recommend and fresh relevant recipe at family.
Again for example, the degree of association of alternating content and interest content in generic feature can be analyzed.The classification is special
Levy the classification or classification for being primarily referred to as belonging to content.For example, in take-away system, dessert belongs to afternoon tea classification, drink also belongs to
In afternoon tea classification, both have certain degree of association (being here identical) on category feature.Based on this, if targeted customer purchases
Dessert is bought, drink can have been recommended to targeted customer.
Again for example, the degree of association of alternating content and interest content on affiliated theme feature can be analyzed.The theme is special
Levy subject scenes or the type of theme for being primarily referred to as that content is related to.For example, by taking the dining room in take-away system as an example, may relate to
Virgin theme dining room, adult's theme dining room, Cantonese teahouse, trendy styles from Hong Kong teahouse etc..Again for example, with the commodity in e-commerce field
As a example by, may relate to wedding celebration theme, outdoor theme, household electrical appliances theme etc..
Again for example, the degree of association of alternating content and interest content in details feature can be analyzed.The details feature master
Refer to the details of content.
In above-described embodiment or following embodiments, analysis alternating content can be with the feature association degree of interest content：
From interest content, the interest characteristics set of targeted customer is extracted；For each interest characteristics in interest characteristics set configures power
Weight；The interest characteristics associated with alternating content in analysis interest characteristics set；According to the interest characteristics associated with alternating content
Weight, obtains feature association degree.
On the basis of features described above correlation analysis scheme, another embodiment of the application provides a kind of commending contents side
Method, its flow as shown in Fig. 2 including：
201st, from the multiple network behavioral data of targeted customer, the interest content of targeted customer is extracted.
202nd, from interest content, the interest characteristics set of targeted customer is extracted.
203rd, for each interest characteristics in interest characteristics set configures weight.
204th, the interest characteristics associated with alternating content in analysis interest characteristics set.
205th, according to the weight of the interest characteristics associated with alternating content, obtain alternating content and closed with the feature of interest content
Connection degree.
206th, according to feature association degree, interest-degree of the targeted customer to alternating content is determined.
207th, according to interest-degree, the selection target content from alternating content.
208th, object content is recommended to targeted customer.
Referring to step 201, the multiple network behavioral data of targeted customer can be obtained, therefrom extract the interest of targeted customer
Content, specifically describes and can be found in previous embodiment, will not be repeated here.
With continued reference to step 202, based on the interest content that step 201 is extracted, the interest characteristics collection of targeted customer is extracted
Close.
Optionally, in the present embodiment or following embodiments, individually from each interest content can extract target and use
The interest characteristics at family, forms the interest characteristics set of targeted customer.Wherein it is possible to the feature that interest content has is extracted, as
The interest characteristics of targeted customer.Illustrate：Assuming that the interest content of targeted customer includes mousse cake and health care of food class hundred
Section, then can extract the interest characteristics of mousse, cake or dessert as targeted customer from mousse cake this interest content,
The interest characteristics of health products or health care as targeted customer can be extracted from health care of food class encyclopaedia this interest content, from
And formation includes the interest characteristics set of mousse, cake or dessert, health products or health care.
Optionally, in the present embodiment or following embodiments, associatedly from interest content can extract targeted customer's
Interest characteristics, forms the interest characteristics set of targeted customer.For example, the mode for associatedly extracting interest characteristics set can be with
It is：According to certain standard, by interest content classification；For the interest content in same class, it is extracted in such interest content
Existing frequency highest feature, as the interest characteristics of targeted customer.Again for example, the side for associatedly extracting interest characteristics set
Formula can be：The frequency for occurring in interest content from the feature for extracting each interest content, the feature for counting extracted respectively
Rate, meets the requirements (for example frequency is more than predetermined threshold value, or frequency ranking is before prescribed percentage) according to frequency selection
Feature, as the interest characteristics of targeted customer.Illustrate：Assuming that the interest content of targeted customer includes that baby is defervescence plaster used, youngster
Virgin health knowledge encyclopaedia, kids' recipe, health-caring equipment etc., then can associatedly analyze these interest contents, therefrom extract children,
Medicine, health care etc. are used as targeted customer's feature interested.
It is the interest characteristics configuration weight in the interest characteristics set that step 202 is extracted with continued reference to step 203.Can
Choosing, be that interest characteristics configuration weight can be in the present embodiment or following embodiments：Obtain the click of the interest characteristics
Amount；It is interest characteristics configuration weight according to its click volume.In general, the click volume of interest characteristics is higher, and its weight is bigger.
Optionally, the click volume of interest characteristics can be click volume of the interest characteristics in application platform, and application platform can directly unite
Count out the click volume of interest characteristics.Or, the click volume of interest characteristics can be interest characteristics click volume on a search engine.
In this case, the interface that can be provided by search engine, obtains the click volume of the interest characteristics from search engine.
With continued reference to step 204, based on the interest characteristics set that step 202 is extracted, what analysis was associated with alternating content
Interest characteristics.
Optionally, in the present embodiment or following embodiments, will can have with alternating content in interest characteristics set
Feature identical interest characteristics, as the interest characteristics associated with alternating content.For example, interest characteristics set can be obtained one by one
In interest characteristics；Judge whether acquired interest characteristics belongs to the feature that alternating content has；If belonged to, it is determined that institute
The interest characteristics of acquisition is the interest characteristics associated with alternating content.Or, the feature in alternating content can be extracted, as time
Select feature；Candidate feature is matched in interest characteristics set；Obtain candidate feature matching in interest characteristics, as with candidate
The interest characteristics of relevance.
Optionally, in the present embodiment or following embodiments, will can also have with alternating content in interest characteristics set
Feature there is the interest characteristics of incidence relation, as the feature related to alternating content.For example, it is assumed that in interest characteristics set
Certain interest characteristics be food materials, alternating content has the characteristic that recipe, then food materials have incidence relation with recipe, so food materials
Belong to the interest characteristics related to alternating content.
Optionally, in the present embodiment or following embodiments, can also analyze in interest characteristics set each interest characteristics with
Similarity between the feature that alternating content has, the interest characteristics using similarity higher than threshold value is used as related to alternating content
Interest characteristics.For example, it is assumed that certain interest characteristics in interest characteristics set is import snacks, alternating content has the characteristic that small
Snacks, then import snacks are high with the similarity of small snacks, so import snacks belong to the interest characteristics related to alternating content.
Illustrate herein, the present embodiment does not limit the execution sequence of above-mentioned steps 203 and step 204, except first carrying out step
Rapid 203 perform outside step 204 again, with parallel execution of steps 203 and step 204, or can also first carry out step
204, then perform step 203.
With continued reference to step 205, associated with alternating content based on what the weight and step 204 configured in step 203 were obtained
Interest characteristics, calculate the feature association degree of alternating content and interest content.
Optionally, in the present embodiment or following embodiments, obtaining feature association degree can be：Will be related to alternating content
The weight of interest characteristics be added, will add up feature association degree of the result as alternating content and interest content, but be not limited to
It is added this mode.For example, it is also possible to using the weight of the interest characteristics related to alternating content as default polynomial coefficient,
The feature degree of correlation is obtained according to default polynomial computation.
With continued reference to step 206, based on the feature association degree that step 205 is obtained, determine targeted customer to alternating content
Interest-degree.Optionally, in the present embodiment or other embodiments, according to feature association degree, determine targeted customer to alternating content
Interest-degree, Ke Yishi：By feature association degree, directly as targeted customer to the interest-degree of alternating content；Or, feature is closed
Connection degree is multiplied by the corresponding coefficient of alternating content, as targeted customer to the interest-degree of alternating content；Or make feature association degree
It is the parameter (such as index) in the corresponding computing formula of alternating content, targeted customer is calculated in candidate according to computing formula
Interest-degree of appearance, etc..
With continued reference to step 207 and 208, based on the interest-degree that step 206 determines, select to recommend from alternating content
To the content (i.e. object content) of targeted customer, and object content is recommended into targeted customer.
Optionally, in the present embodiment or other embodiments, mesh can be selected from alternating content according to interest-degree just
Mark content.For example, the interest-degree of alternating content and threshold value can be compared, interest-degree is made higher than the alternating content of threshold value
It is object content.Or, ranking can be carried out to alternating content according to interest-degree order from high to low, obtain ranking and most lean on
One or more preceding alternating contents are used as object content.Or, can be according to interest-degree order from high to low, in candidate
Appearance carries out ranking, and the alternating content for choosing certain proportion (such as preceding 5 percent, 1 15 etc.) according to rank order is made
It is object content.Or, the alternating content that interest-degree is located in specified range can be selected as in target according to interest-degree
Hold.
Optionally, in the present embodiment or other embodiments, in order to the interest of more comprehensive coverage goal user
Point, according to interest-degree, from alternating content during selection target content, can be with diversification ground selection target content.Institute
Diversification is stated to be primarily referred to as while interest-degree requirement is met, different classes of content being selected as far as possible.So can be to user
Recommend various different classes of contents, can more meet the demand of targeted customer, improve targeted customer to application platform
Experience Degree, be conducive to exciting targeted customer that application spontaneously is recommended into the social activity of oneself, increase the customer flow of application.
Optionally, the object content based on above-mentioned Multiple strategies, it is considered to which the way of realization of different content is different, for letter
Change and recommend operation, different target content can respectively be recommended by targeted customer by single information or the page.
Optionally, the object content based on above-mentioned Multiple strategies, causes to do in order to avoid repeatedly recommending targeted customer
Disturb, the object content of the Multiple strategies can be recommended to targeted customer with polymerization methodses.For example, diversification can be selected
The object content selected is aggregated in same information or the page, recommends targeted customer.
As can be seen here, interest of the above-described embodiment from user so that the content for a purpose recommended is stronger, meets use
The probability of family demand is higher, can improve user to recommending the Experience Degree of behavior, is conducive to exciting user interest, reaches recommendation mesh
's.
In above-described embodiment or following embodiments, object content comes from alternating content, it is therefore desirable to determine in candidate
Hold.
In an optional embodiment, can directly by the content in platform database, as alternating content.I.e. based on flat
Content in platform database is done and is recommended, and this mode is realized simply, recommending efficiency higher.
In another optional embodiment, alternating content can be determined according to similar users content interested.It is based on
Similar users are done and are recommended, and are conducive to recommending to more conform to the content of its demand to user, improve the accuracy recommended.
In another alternative embodiment mode, can according to application development situation, with reference to platform database and similar users,
Determine alternating content.For example, in application initial stage (i.e. cold-start phase), customer volume is relatively fewer, can be based on platform database
In content do and recommend；When customer volume is accumulated to a certain extent, can be done based on similar users and recommended.By above two scheme knot
Conjunction is used, and is conducive to improving the flexibility of suggested design, can more targetedly to user's content recommendation so that recommended
Content more conform to user's request, be conducive to exciting the interest of user.
During the above-mentioned content interested based on similar users carries out commending contents, can analyze user set in
Similarity between each user content interested and the interest content of targeted customer；According to similarity, obtained from user's set
Take the similar users of targeted customer.
Each user in gathering user, obtains the network behavior data of the user, from the network behavior number of the user
In, user content interested is extracted.Wherein, the process of user's content interested is extracted, with extraction targeted customer's
The process of interest content is identical, reference can be made to the description of previous embodiment.A kind of implementation method, can be by the interest of targeted customer
Hold the content interested with each user directly to compare, obtain similarity.Another implementation method, can analyze the emerging of targeted customer
Interesting content and each user content interested degree of association characteristically, the degree of association on feature based determine similarity.Root
According to similarity, the similar users of targeted customer are determined from user's set.
Based on similar users content interested, alternating content is determined.Optionally, can to extract similar users interested
Feature in content, as similar features；From interest content, the interest characteristics set of targeted customer is extracted；By similar features
It is compared with interest characteristics set, the distinguishing characteristics of interest characteristics set is not belonging in acquisition similar features；From similar users
Network behavior data or platform database in, obtain with the distinguishing characteristics content, as alternating content.In these candidates
Appearance is similar users interested, can be recommended to targeted customer.
As a example by promoting and take out class application, the customer volume of class application is taken out to expand, can be excited by recommendation function
The interest of user, guiding user actively forwards in the social platform of oneself and takes out the content that class application is recommended, to take out class
Using bringing customer flow.
With reference to Fig. 3, promoting the commending contents process in taking out class application process, including：
With reference to Fig. 3, application initial stage：Customer volume is relatively fewer, using the content in platform database as alternating content.
In addition, determining targeted customer to be recommended, obtain network behavior data of the targeted customer within a period of time recently, such as it is clear
Look at, pay close attention to, placing an order, plus the data that produce of the behavior such as shopping cart.On the one hand, the network behavior data based on targeted customer, extract
The interest content of targeted customer, from interest content, extracts the interest characteristics set of targeted customer, and according to interest characteristics set
In click volume of each interest characteristics on search engine or application platform, be each interest characteristics configuration weight；On the other hand, from time
Select and extract candidate feature in content.Calculate the similarity between interest characteristics set and candidate feature；Commending contents：Based on the phase
Commending contents are carried out like degree.
With continued reference to Fig. 3, application maturity period：The customer volume accumulation of class application is taken out to a certain extent, according to user
Each user content interested and the interest content of targeted customer, determine the similar users of targeted customer in set；Based on similar
User's content interested, from similar users content interested, extracts similar features；Calculate similar features and interest characteristics
Similarity between set；Commending contents are carried out based on similarity.
In the above-described embodiments, according to the popularization situation for taking out class application, pushed away with reference to platform database and similar users
Recommend, in each period, can recommended user's content interested, be conducive to improve user to take out class application Experience Degree, have
Recommend to take out class application to other users beneficial to user's active, or forwarded in social platform in the application recommendation of take-away class
Hold, and then increase the customer flow for taking out class application, reach recommendation purpose.
It should be noted that the executive agent that above-described embodiment provides each step of method may each be same equipment,
Or, the method is also by distinct device as executive agent.Such as, step 101 to the executive agent of step 104 can be equipment
A；Again such as, step 101 and 102 executive agent can be device A, and step 103 and 104 executive agent can be equipment B；
Etc..
The structural representation of the content recommendation device that Fig. 4 is provided for the another embodiment of the application.As shown in figure 4, the device
Including：Extraction module 41, analysis module 42, selecting module 43 and recommending module 44.
Extraction module 41, for from the multiple network behavioral data of targeted customer, extracting the interest content of targeted customer.
Analysis module 42, for the interest content extracted according to extraction module 41, targeted customer is to alternating content for analysis
Interest-degree.
Selecting module 43, for the interest-degree analyzed according to analysis module 42, the selection target content from alternating content.
Recommending module 44, for recommending the selected object content of selecting module 43 to targeted customer.
In an optional embodiment, as shown in figure 5, analysis module 42 includes：Analytic unit 421 and determining unit 422.
Analytic unit 421, the feature association degree for analyzing the interest content that alternating content and extraction module 41 are extracted.
Determining unit 422, for the feature association degree analyzed according to analytic unit 421, determines targeted customer to candidate
The interest-degree of content.
It is further alternative, analytic unit 421 specifically for：
From interest content, the interest characteristics set of targeted customer is extracted；
For each interest characteristics in interest characteristics set configures weight；
The interest characteristics associated with alternating content in analysis interest characteristics set；
According to the weight of the interest characteristics associated with alternating content, the feature association degree is obtained.
It is further alternative, selecting module 43 specifically for：According to interest-degree, from alternating content, the selection of diversification ground
Object content.Correspondingly, recommending module 44 specifically for：With polymerization methodses, the diversification of selecting module 43 is recommended to targeted customer
The object content of selection.
As shown in figure 5, the device also includes：Determining module 45, for the content in direct access platform database, as
Alternating content；And/or, similar users according to targeted customer content interested determines alternating content.
It is further alternative, determining module 45 according to similar users content interested, when determining alternating content, specifically
For：
The feature in similar users content interested is extracted, as similar features；
From interest content, the interest characteristics set of targeted customer is extracted；
The distinguishing characteristics of interest characteristics set is not belonging in acquisition similar features；
From the network behavior data or platform database of similar users, the content with distinguishing characteristics is obtained, as time
Select content.
The content recommendation device that the present embodiment is provided, can be used to perform the method that above-described embodiment is provided, and relevant content can
Referring to the description of embodiment of the method, will not be repeated here.
The content recommendation device that the present embodiment is provided, can be based on the multiple network behavioral data of user, determine that user's sense is emerging
The content of interest；And then, based on user's content interested, go to analyze interest-degree of the user to alternating content；Based on user to waiting
The interest-degree of content is selected, to user's content recommendation.Interest of the content recommendation device that the present embodiment is provided from user so that
The content for a purpose recommended is stronger, and the probability for meeting user's request is higher, can improve user to recommending the Experience Degree of behavior,
Be conducive to exciting user interest, reach recommendation purpose.
It should be understood by those skilled in the art that, embodiments of the invention can be provided as method, system or computer program
Product.Therefore, the present invention can be using the reality in terms of complete hardware embodiment, complete software embodiment or combination software and hardware
Apply the form of example.And, the present invention can be used and wherein include the computer of computer usable program code at one or more
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) is produced
The form of product.
The present invention is the flow with reference to method according to embodiments of the present invention, equipment (system) and computer program product
Figure and/or block diagram are described.It should be understood that every first-class during flow chart and/or block diagram can be realized by computer program instructions
The combination of flow and/or square frame in journey and/or square frame and flow chart and/or block diagram.These computer programs can be provided
The processor of all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing devices is instructed to produce
A raw machine so that produced for reality by the instruction of computer or the computing device of other programmable data processing devices
The device of the function of being specified in present one flow of flow chart or multiple one square frame of flow and/or block diagram or multiple square frames.
These computer program instructions may be alternatively stored in can guide computer or other programmable data processing devices with spy
In determining the computer-readable memory that mode works so that instruction of the storage in the computer-readable memory is produced and include finger
Make the manufacture of device, the command device realize in one flow of flow chart or multiple one square frame of flow and/or block diagram or
The function of being specified in multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing devices so that in meter
Series of operation steps is performed on calculation machine or other programmable devices to produce computer implemented treatment, so as in computer or
The instruction performed on other programmable devices is provided for realizing in one flow of flow chart or multiple flows and/or block diagram one
The step of function of being specified in individual square frame or multiple square frames.
In a typical configuration, computing device includes one or more processors (CPU), input/output interface, net
Network interface and internal memory.
Internal memory potentially includes the volatile memory in computer-readable medium, random access memory (RAM) and/or
The forms such as Nonvolatile memory, such as read-only storage (ROM) or flash memory (flash RAM).Internal memory is computer-readable medium
Example.
Computer-readable medium includes that permanent and non-permanent, removable and non-removable media can be by any method
Or technology realizes information Store.Information can be computer-readable instruction, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase transition internal memory (PRAM), static RAM (SRAM), moves
State random access memory (DRAM), other kinds of random access memory (RAM), read-only storage (ROM), electric erasable
Programmable read only memory (EEPROM), fast flash memory bank or other memory techniques, read-only optical disc read-only storage (CD-ROM),
Digital versatile disc (DVD) or other optical storages, magnetic cassette tape, the storage of tape magnetic rigid disk or other magnetic storage apparatus
Or any other non-transmission medium, can be used to store the information that can be accessed by a computing device.Defined according to herein, calculated
Machine computer-readable recording medium does not include temporary computer readable media (transitory media), such as data-signal and carrier wave of modulation.
Also, it should be noted that term " including ", "comprising" or its any other variant be intended to nonexcludability
Comprising so that process, method, commodity or equipment including a series of key elements not only include those key elements, but also wrapping
Include other key elements being not expressly set out, or also include for this process, method, commodity or equipment is intrinsic wants
Element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that wanted including described
Also there is other identical element in process, method, commodity or the equipment of element.
It will be understood by those skilled in the art that embodiments herein can be provided as method, system or computer program product.
Therefore, the application can be using the embodiment in terms of complete hardware embodiment, complete software embodiment or combination software and hardware
Form.And, the application can be used to be can use in one or more computers for wherein including computer usable program code and deposited
The shape of the computer program product implemented on storage media (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
Formula.
Embodiments herein is the foregoing is only, the application is not limited to.For those skilled in the art
For, the application can have various modifications and variations.It is all any modifications made within spirit herein and principle, equivalent
Replace, improve etc., within the scope of should be included in claims hereof.
Claims (17)
1. a kind of content recommendation method, it is characterised in that including：
From the multiple network behavioral data of targeted customer, the interest content of the targeted customer is extracted；
According to the interest content, interest-degree of the targeted customer to alternating content is analyzed；
According to the interest-degree, the selection target content from the alternating content；
Recommend the object content to the targeted customer.
2. method according to claim 1, it is characterised in that the analytical procedure of the interest-degree, including：
Analyze the feature association degree of the alternating content and the interest content；
According to the feature association degree, interest-degree of the targeted customer to the alternating content is determined.
3. method according to claim 2, it is characterised in that the analytical procedure of the feature association degree, including：
From the interest content, the interest characteristics set of the targeted customer is extracted；
For each interest characteristics in the interest characteristics set configures weight；
Analyze the interest characteristics associated with the alternating content in the interest characteristics set；
According to the weight of the interest characteristics associated with the alternating content, the feature association degree is obtained.
4. method according to claim 3, it is characterised in that the configuration step of the weight, including：
It is the interest characteristics configuration weight according to the click volume of the interest characteristics.
5. method according to claim 3, it is characterised in that the interest characteristics associated with the alternating content point
Analysis step, including：
The feature in the alternating content is extracted, as candidate feature；
The candidate feature is matched in the interest characteristics set；
Obtain the interest characteristics in the candidate feature matching.
6. the method according to claim any one of 1-5, it is characterised in that the selection step of the object content, including：
According to the interest-degree, from the alternating content, the diversification ground selection object content.
7. method according to claim 6, it is characterised in that the recommendation step of the object content, including：
With polymerization methodses, the object content of the Multiple strategies is recommended to the targeted customer.
8. the method according to claim any one of 1-5, it is characterised in that before the interest-degree is analyzed, also include：
Directly by the content in platform database, as the alternating content；And/or
Similar users according to targeted customer content interested, determines the alternating content.
9. method according to claim 8, it is characterised in that also include：
Similarity in analysis user's set between each user content interested and the interest content；
According to the similarity, the similar users of the targeted customer are obtained from user set.
10. method according to claim 8, it is characterised in that the determination step of the alternating content, including：
The feature in similar users content interested is extracted, as similar features；
From the interest content, the interest characteristics set of the targeted customer is extracted；
Obtain the distinguishing characteristics that the interest characteristics set is not belonging in the similar features；
From the network behavior data or platform database of the similar users, the content with the distinguishing characteristics is obtained, made
It is the alternating content.
A kind of 11. content recommendation devices, it is characterised in that including：
Extraction module, for from the multiple network behavioral data of targeted customer, extracting the interest content of the targeted customer；
Analysis module, for according to the interest content, analyzing interest-degree of the targeted customer to alternating content；
Selecting module, for according to the interest-degree, the selection target content from the alternating content；
Recommending module, for recommending the object content to the targeted customer.
12. devices according to claim 11, it is characterised in that the analysis module includes：
Analytic unit, the feature association degree for analyzing the alternating content and the interest content；
Determining unit, for according to the feature association degree, determining interest-degree of the targeted customer to the alternating content.
13. devices according to claim 11, it is characterised in that the analytic unit specifically for：
From the interest content, the interest characteristics set of the targeted customer is extracted；
For each interest characteristics in the interest characteristics set configures weight；
Analyze the interest characteristics associated with the alternating content in the interest characteristics set；
According to the weight of the interest characteristics associated with the alternating content, the feature association degree is obtained.
14. device according to claim any one of 11-13, it is characterised in that the selecting module specifically for：
According to the interest-degree, from the alternating content, the diversification ground selection object content.
15. devices according to claim 14, it is characterised in that the recommending module specifically for：
With polymerization methodses, the object content of the Multiple strategies is recommended to the targeted customer.
16. device according to claim any one of 11-13, it is characterised in that also include：
Determining module, for the content in direct access platform database, as the alternating content；And/or, according to the mesh
The similar users of user content interested is marked, the alternating content is determined.
17. devices according to claim 16, it is characterised in that the determining module specifically for：
The feature in similar users content interested is extracted, as similar features；
From the interest content, the interest characteristics set of the targeted customer is extracted；
Obtain the distinguishing characteristics that the interest characteristics set is not belonging in the similar features；
From the network behavior data or platform database of the similar users, the content with the distinguishing characteristics is obtained, made
It is the alternating content.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610997930.8A CN106776697A (en) | 2016-11-11 | 2016-11-11 | Content recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610997930.8A CN106776697A (en) | 2016-11-11 | 2016-11-11 | Content recommendation method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106776697A true CN106776697A (en) | 2017-05-31 |
Family
ID=58973525
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610997930.8A Pending CN106776697A (en) | 2016-11-11 | 2016-11-11 | Content recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106776697A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107392704A (en) * | 2017-07-13 | 2017-11-24 | 北京小度信息科技有限公司 | Information recommendation method and device |
| CN109033408A (en) * | 2018-08-03 | 2018-12-18 | 腾讯科技（深圳）有限公司 | Information-pushing method and device, computer readable storage medium, electronic equipment |
| CN109145193A (en) * | 2017-06-27 | 2019-01-04 | 阿里巴巴集团控股有限公司 | A kind of information-pushing method and system |
| CN109446401A (en) * | 2017-08-29 | 2019-03-08 | 合肥美的智能科技有限公司 | User's preferred diet portrait construction method, device and machine readable storage medium |
| CN109618231A (en) * | 2018-11-15 | 2019-04-12 | 深圳市酷开网络科技有限公司 | A kind of movie and television contents recommended method and system based on cold start-up |
| CN112685676A (en) * | 2020-12-30 | 2021-04-20 | 华控清交信息科技(北京)有限公司 | Information recommendation method and device and electronic equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102929959A (en) * | 2012-10-10 | 2013-02-13 | 杭州东信北邮信息技术有限公司 | Book recommendation method based on user actions |
| CN104008195A (en) * | 2014-06-13 | 2014-08-27 | 清华大学 | Service recommendation method oriented to service combination |
| CN104036038A (en) * | 2014-06-30 | 2014-09-10 | 北京奇虎科技有限公司 | News recommendation method and system |
| CN104572863A (en) * | 2014-12-19 | 2015-04-29 | 阳珍秀 | Product recommending method and system |
| CN105488233A (en) * | 2016-01-25 | 2016-04-13 | 广东顺德中山大学卡内基梅隆大学国际联合研究院 | Reading information recommendation method and system |
| 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 |
| CN106227870A (en) * | 2016-07-28 | 2016-12-14 | 乐视控股（北京）有限公司 | A kind of video website recommends method and device |
-
2016
- 2016-11-11 CN CN201610997930.8A patent/CN106776697A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102929959A (en) * | 2012-10-10 | 2013-02-13 | 杭州东信北邮信息技术有限公司 | Book recommendation method based on user actions |
| CN104008195A (en) * | 2014-06-13 | 2014-08-27 | 清华大学 | Service recommendation method oriented to service combination |
| CN104036038A (en) * | 2014-06-30 | 2014-09-10 | 北京奇虎科技有限公司 | News recommendation method and system |
| CN104572863A (en) * | 2014-12-19 | 2015-04-29 | 阳珍秀 | Product recommending method and system |
| 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 |
| CN105488233A (en) * | 2016-01-25 | 2016-04-13 | 广东顺德中山大学卡内基梅隆大学国际联合研究院 | Reading information recommendation method and system |
| CN106227870A (en) * | 2016-07-28 | 2016-12-14 | 乐视控股（北京）有限公司 | A kind of video website recommends method and device |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109145193A (en) * | 2017-06-27 | 2019-01-04 | 阿里巴巴集团控股有限公司 | A kind of information-pushing method and system |
| CN107392704A (en) * | 2017-07-13 | 2017-11-24 | 北京小度信息科技有限公司 | Information recommendation method and device |
| CN109446401A (en) * | 2017-08-29 | 2019-03-08 | 合肥美的智能科技有限公司 | User's preferred diet portrait construction method, device and machine readable storage medium |
| CN109033408A (en) * | 2018-08-03 | 2018-12-18 | 腾讯科技（深圳）有限公司 | Information-pushing method and device, computer readable storage medium, electronic equipment |
| CN109033408B (en) * | 2018-08-03 | 2021-10-22 | 腾讯科技（深圳）有限公司 | Information pushing method and device, computer readable storage medium and electronic equipment |
| CN109618231A (en) * | 2018-11-15 | 2019-04-12 | 深圳市酷开网络科技有限公司 | A kind of movie and television contents recommended method and system based on cold start-up |
| CN109618231B (en) * | 2018-11-15 | 2021-07-27 | 深圳市酷开网络科技股份有限公司 | Movie content recommendation method and system based on cold start |
| CN112685676A (en) * | 2020-12-30 | 2021-04-20 | 华控清交信息科技(北京)有限公司 | Information recommendation method and device and electronic equipment |
| CN112685676B (en) * | 2020-12-30 | 2023-12-15 | 华控清交信息科技(北京)有限公司 | Information recommendation method and device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106776697A (en) | Content recommendation method and device | |
| KR101735602B1 (en) | Personalized recommendation method, system and computer-readable record medium | |
| CN103823908B (en) | Content recommendation method and server based on user preference | |
| US9691096B1 (en) | Identifying item recommendations through recognized navigational patterns | |
| CN108960992A (en) | A kind of information recommendation method and relevant device | |
| JP6288683B2 (en) | User profiles based on clustered hierarchy descriptors | |
| US8838521B2 (en) | Systems and methods for trend aware self-correcting entity relationship extraction | |
| US20140081800A1 (en) | Recommending Product Information | |
| CN111984837B (en) | Commodity data processing method, device and equipment | |
| CN110335123B (en) | Commodity recommendation method, system, computer readable medium and device based on social e-commerce platform | |
| CN109447713A (en) | A kind of recommended method and device of knowledge based map | |
| CN109087177A (en) | To the method, apparatus and computer readable storage medium of target user's Recommendations | |
| CN106202516A (en) | A kind of e-commerce platform merchandise display method according to timing node | |
| CN103646341B (en) | A kind of website provides the recommendation method and apparatus of object | |
| US9767417B1 (en) | Category predictions for user behavior | |
| CN109493113A (en) | A kind of providing method and device of favor information | |
| US9767204B1 (en) | Category predictions identifying a search frequency | |
| CN107248095A (en) | Recommend method and device | |
| CN104809637A (en) | Commodity recommending method and system realized by computer | |
| CN110933472B (en) | Method and device for realizing video recommendation | |
| KR102097045B1 (en) | Method and apparatus to recommend products reflecting characteristics of users | |
| US10474670B1 (en) | Category predictions with browse node probabilities | |
| CN105027123A (en) | Recommending content based on proxy-based preference indications | |
| US10282474B2 (en) | Category constrained queries | |
| KR20150024465A (en) | System and method for recommending contents based on user characteristics and transactions, and recording media thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 100085 Beijing, Haidian District on the road to the information on the ground floor of the 1 to the 3 floor of the 2 floor, room 11, 202 Applicant after: Beijing Xingxuan Technology Co.,Ltd. Address before: 100085 Beijing, Haidian District on the road to the information on the ground floor of the 1 to the 3 floor of the 2 floor, room 11, 202 Applicant before: Beijing Xiaodu Information Technology Co.,Ltd. |