CN106354858A - Information resource recommendation method based on label clusters - Google Patents
Information resource recommendation method based on label clusters Download PDFInfo
- Publication number
- CN106354858A CN106354858A CN201610804298.0A CN201610804298A CN106354858A CN 106354858 A CN106354858 A CN 106354858A CN 201610804298 A CN201610804298 A CN 201610804298A CN 106354858 A CN106354858 A CN 106354858A
- Authority
- CN
- China
- Prior art keywords
- user
- information
- knowledge
- label
- recommendation
- 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)
Abstract
The invention discloses an information resource recommendation method based on label clusters. The method mainly includes the steps that knowledge owned by single users is clustered, and different interest fields of the single users are distinguished; clustering is conducted on the knowledge of all users according to label cloud corresponding to different interest fields, and different knowledge themes are formed, wherein the knowledge themes are clusters of information resources or user clusters with same theme labels; by means of the collaborative filtering mode, resources which the same kind of users pay attention to are recommended to target users. Thus, the advantages that accuracy of recommendation is improved, the recommendation effect is improved, adhesiveness of the users is improved, interaction operation of people and the internet is achieved, the operation mode that information actively serves people is implemented, and social benefits and economic benefits are improved can be achieved.
Description
Technical field
A kind of the present invention relates to information resources recommend method and technology field, in particular it relates to information based on label clustering
Resource recommendation method.
Background technology
With the arrival in web2.0 epoch, the various media information service terminal such as broadcasting and TV, the Internet, mobile network is daily all
The information resources that magnanimity can be produced present in front of the user, and we enter the so-called big data epoch.Meanwhile, now
Most people live allegro " fast food type " life, in the face of huge information ocean, how therefrom to find and meet user certainly
Body interest just becomes quite to be difficult with the information needing, thus the problems such as create " information overflow " and " information dense fog ".
Traditional search engine cannot meet needs, although because this search engine technique achieves to a certain extent
Information filtering, but the characteristic of user is not taken into account by it, all users of equivalent processes, returns same resource sequence to use
Family, and quantity of information is quite big.Therefore, how to be found from internet mass information according to the personal preference of user and meet its demand
Information, and then recommend user, it has also become the problem of this field relevant personnel concern all the more in recent years.Personalized recommendation technology
Also arise at the historic moment.
Wide proposed algorithm is applied to have content-based recommendation algorithm, collaborative filtering, mixing proposed algorithm etc..
Collaborative filtering be current technology the most ripe be also most widely used recommended technology, but information resources are varied, single
Proposed algorithm just seem awkward in actual system application, how to recommend method digging user emerging using suitable
Accurate recommendation that is interesting, realizing resource remains current a great problem.
During realizing the present invention, inventor finds at least to exist in prior art does not have suitable method to user
Interest is excavated, the defect such as the recommendation of information resources is not accurate.
Content of the invention
It is an object of the invention to, for the problems referred to above, propose a kind of information resources based on label clustering and recommend method,
To realize quick and precisely digging user interest, realize the accurate recommendation of information resources, the advantage of lifting recommendation effect.
For achieving the above object, the technical solution used in the present invention is: a kind of information resources based on label clustering are recommended
Method, specifically includes that
Step 1: the knowledge that unique user is had is clustered, distinguishes the different interest worlds of unique user；
Step 2: be to cluster according to the knowledge of all users with the corresponding label-cloud in different interest worlds, formed different
Knowledget opic, described knowledget opic is the cluster of information resources or has the user clustering with same subject label；
Step 3: using collaborative filtering mode, by the resource recommendation of same knowledget opic user concern to targeted customer.
Further, step 3 includes:
Step 31: judge, with the intensity of knowledget opic user-to-user information interactive relationship, to select information interaction relationship strength
High a pair of user；
Step 32: need according to the information that the knowledge belonging to this knowledget opic of one of user annotation predicts another user
Ask, produce the information recommendation object for another user；
Step 33: judge whether that a user corresponds to two knowledget opics, if existing, this user constitutes two knowledge masters
It is given in the second knowledget opic in the information recommendation that the first knowledget opic marks by " bridge " and interacts therewith relation by " bridge " of topic
Another user the strongest, realizes that information resources are inhomogeneous transboundary to flow, and knows when the information resources flowing across theme enter second
After knowing theme, another user of the second knowledge passes through step 31 and step 32 by information recommendation to other in the second knowledget opic
User.
Further, described information resource includes text, picture, audio or video.
A kind of information resources based on label clustering of various embodiments of the present invention recommend method, due to specifically including that list
The knowledge that individual user has is clustered, and distinguishes the different interest worlds of unique user；With the corresponding label in different interest worlds
Cloud is that foundation clusters to the knowledge of all users, forms different knowledget opics, described knowledget opic is information resources
Cluster or have the user clustering with same subject label；Using collaborative filtering mode, the resource that fellow users are paid close attention to
Recommend targeted customer；Such that it is able to realize improving the accuracy recommended, lift recommendation effect, increase the cohesiveness of user, real
The interactive operation of existing people and the Internet, the information of realization is taken the initiative in offering a hand operating mode, the social benefit of raising and the economic effect of the mankind
The advantage of benefit.
Other features and advantages of the present invention will illustrate in the following description, and, partly become from description
Obtain it is clear that or being understood by implementing the present invention.
Below by drawings and Examples, technical scheme is described in further detail.
Brief description
Accompanying drawing is used for providing a further understanding of the present invention, and constitutes a part for description, the reality with the present invention
Apply example and be used for explaining the present invention together, be not construed as limiting the invention.In the accompanying drawings:
Fig. 1 is that the information resources based on label clustering described in the embodiment of the present invention recommend being formed based on label clustering of method to use
Family correlation network figure；
Fig. 2 is the combination collaborative filtering thought pair of the information resources recommendation method based on label clustering described in the embodiment of the present invention
Cluster user is realized recommending schematic diagram.
Specific embodiment
Below in conjunction with accompanying drawing, the preferred embodiments of the present invention are illustrated it will be appreciated that preferred reality described herein
Apply example to be merely to illustrate and explain the present invention, be not intended to limit the present invention.
Specifically, a kind of information resources based on label clustering recommend method, specifically include that
Step 1: the knowledge that unique user is had is clustered, distinguishes the different interest worlds of unique user；
Step 2: be to cluster according to the knowledge of all users with the corresponding label-cloud in different interest worlds, formed different
Knowledget opic, described knowledget opic is the cluster of information resources or has the user clustering with same subject label；
Step 3: using collaborative filtering mode, by the resource recommendation of same knowledget opic user concern to targeted customer.
Step 3 includes:
Step 31: judge, with the intensity of knowledget opic user-to-user information interactive relationship, to select information interaction relationship strength
High a pair of user；
Step 32: need according to the information that the knowledge belonging to this knowledget opic of one of user annotation predicts another user
Ask, produce the information recommendation object for another user；
Step 33: judge whether that a user corresponds to two knowledget opics, if existing, this user constitutes two knowledge masters
It is given in the second knowledget opic in the information recommendation that the first knowledget opic marks by " bridge " and interacts therewith relation by " bridge " of topic
Another user the strongest, realizes that information resources are inhomogeneous transboundary to flow, and knows when the information resources flowing across theme enter second
After knowing theme, another user of the second knowledge passes through step 31 and step 32 by information recommendation to other in the second knowledget opic
User.
Described information resource includes text, picture, audio or video
Label has relatedness, cluster property, Textuality, can be by the change of network hotspot and personal focus by label record
Fully demonstrate out, the relation of contact user and label, use the labels as medium, not only can more easily excavate user's
Preference and custom, are also more convenient the user having common interest hobby implicitly condenses together.Therefore can in conjunction with cluster and
Traditional collaborative filtering, realizes the information resources Collaborative Recommendation based on label clustering using label.
Those are had the user clustering of identical interest, similar thinking and expression way, between them, produces information
Interactive probability is relatively large, such that it is able to make up the deficiency of mass society network.
If according to traditional analytical mathematics, the user clustering based on tag uses the similar of label it is only necessary to compare user
Property just can be found that those have the user group of similar tags use pattern.However, due to the diversity of user interest, each
The corresponding label in interest worlds there may be larger difference, if all labels that direct basis user uses carry out similarity
Analysis, will necessarily be greatly reduced the precision of user clustering.
In conjunction with Fig. 1: clustered the knowledge that unique user is had first, to distinguish the different interest necks of unique user
Domain, is then to cluster according to the knowledge of all users with the corresponding label-cloud in different interest worlds, and then is formed different
Knowledget opic.Obviously, the knowledget opic so obtaining is not only just the cluster of information resources, is also to have these masters simultaneously
The cluster of the user of topic label.
Utilize the thought of collaborative filtering afterwards, the resource recommendation that fellow users are paid close attention to is to targeted customer.With general society
Can network different, in interest topic network, it is interactive that the relation between user embodies each other, is more convenient for recommending to believe accordingly
Breath content is to targeted customer.Simultaneously it is recommended that the information exchange behavior inherently between user will be it is recommended that effect will directly affect use
Family relationship strength each other, can be used for the interest topic network optimization again, and then realizes " interest topic network → information recommendation
The benign cycle of → feedback of the information → interest topic the network optimization → information recommendation ... ".In addition, the relatedness of label, cluster
Property, Textuality can also reduce the sparse sex chromosome mosaicism of data.
In conjunction with Fig. 2, inside each interest topic, collaborative filtering thought can be used for reference and realize information resources recommendation.As
Information interaction relationship strength between fruit user a and user b is higher, can belong to knowing of interest topic 2 according to what user a marked
Know and to predict the information requirement of user b, and then produce the information recommendation object for user b.Meanwhile, between different themes, profit
Realize the information recommendation across interest topic with " bridge ", to make up the deficiency of information recommendation in interest topic.Play " bridge " role
Be exactly the user that those simultaneously appear in different interest topics.In fig. 2, user c simultaneously appears in classification 2 and classification 3,
User c just constructs " bridge " connecting theme 2 and knowledget opic 3.Thus, used by comparing in user c and theme 3 in theme 2
The label of family d mark, by the information recommendation of user c mark in theme 2 to user d, and then can realize information resources from theme
2 flowings transboundary arriving theme 3.Obviously, after this information resources across theme flowing enter theme 3, can be by being based on
The method of collaborative filtering recommends the other users in theme 3.
Present invention achieves: (1) does medium using label, and the recommendation for information resources provides new thinking, by label
Cluster realizes the classification polymerization of interest crowd, in conjunction with traditional collaborative filtering recommending, the user in each class interest is realized
Similar recommendation；
(2) it is likely to that this thinking of common point is had on other interest using the user with a certain same interest, by certain simultaneously
The user also in category of interest with other interest, as " bridge ", excavates the potential interest of this category of interest user；
(3) different types of resource content can be realized with tag extraction work, therefore this recommendation method can be used for recommending literary composition
All kinds of resource information such as basis, picture, audio frequency, video, applied range is it is recommended that easy realization simple to operate is it is recommended that effect is accurately complete
Face.
Following beneficial effect at least can be reached: realize individual character using the cluster of label with reference to collaborative filtering and recommend,
The theme using same label being able to conveniently find, thus finding the user having common hobby and Focus Area, with label being
Medium, makes full use of the advantage of label, can preferably record and embody the change of personal focus, facilitate commending system to find
" fellow traveler ", the relatedness of label, cluster property, Textuality can also reduce the sparse sex chromosome mosaicism of data, improve the accurate of recommendation
Degree, lifts recommendation effect, increases the cohesiveness of user, realizes the interactive operation of people and the Internet, and the information of realization is taken the initiative in offering a hand people
The operating mode of class, has very high Social benefit and economic benefit.
Finally it is noted that the foregoing is only the preferred embodiments of the present invention, it is not limited to the present invention,
Although being described in detail to the present invention with reference to the foregoing embodiments, for a person skilled in the art, it still may be used
To modify to the technical scheme described in foregoing embodiments, or equivalent is carried out to wherein some technical characteristics.
All any modification, equivalent substitution and improvement within the spirit and principles in the present invention, made etc., should be included in the present invention's
Within protection domain.
Claims (3)
1. a kind of information resources based on label clustering recommend method: it is characterized in that including:
Step 1: the knowledge that unique user is had is clustered, distinguishes the different interest worlds of unique user；
Step 2: be to cluster according to the knowledge of all users with the corresponding label-cloud in different interest worlds, formed different
Knowledget opic, described knowledget opic is the cluster of information resources or has the user clustering with same subject label；
Step 3: using collaborative filtering mode, the information resources of same knowledget opic user concern are recommended targeted customer.
2. information resources based on label clustering according to claim 1 recommend method it is characterised in that described step 3
Including:
Step 31: judge the intensity of same knowledget opic user-to-user information interactive relationship, select information interaction relationship strength
High a pair of user；
Step 32: need according to the information that the knowledge belonging to this knowledget opic of one of user annotation predicts another user
Ask, produce the information recommendation object for another user；
Step 33: judge whether that a user corresponds to two knowledget opics, if existing, this user constitutes two knowledge masters
It is given in the second knowledget opic in the information recommendation that the first knowledget opic marks by " bridge " and interacts therewith relation by " bridge " of topic
Another user the strongest, realizes that information resources are inhomogeneous transboundary to flow, and knows when the information resources flowing across theme enter second
After knowing theme, another user of the second knowledge passes through step 31 and step 32 by information recommendation to other in the second knowledget opic
User.
3. information resources based on label clustering according to claim 1 and 2 recommend method it is characterised in that described letter
Breath resource includes text, picture, audio or video.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610804298.0A CN106354858A (en) | 2016-09-06 | 2016-09-06 | Information resource recommendation method based on label clusters |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610804298.0A CN106354858A (en) | 2016-09-06 | 2016-09-06 | Information resource recommendation method based on label clusters |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106354858A true CN106354858A (en) | 2017-01-25 |
Family
ID=57859077
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610804298.0A Pending CN106354858A (en) | 2016-09-06 | 2016-09-06 | Information resource recommendation method based on label clusters |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106354858A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107679883A (en) * | 2017-05-05 | 2018-02-09 | 平安科技（深圳）有限公司 | The method and system of advertisement generation |
| CN107688600A (en) * | 2017-07-12 | 2018-02-13 | 百度在线网络技术（北京）有限公司 | Knowledge point method for digging and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103810192A (en) * | 2012-11-09 | 2014-05-21 | 腾讯科技（深圳）有限公司 | User interest recommending method and device |
| US8825672B1 (en) * | 2010-09-20 | 2014-09-02 | Amazon Technologies, Inc. | System and method for determining originality of data content |
| CN105045931A (en) * | 2015-09-02 | 2015-11-11 | 南京邮电大学 | Video recommendation method and system based on Web mining |
| CN105512326A (en) * | 2015-12-23 | 2016-04-20 | 成都品果科技有限公司 | Picture recommending method and system |
-
2016
- 2016-09-06 CN CN201610804298.0A patent/CN106354858A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8825672B1 (en) * | 2010-09-20 | 2014-09-02 | Amazon Technologies, Inc. | System and method for determining originality of data content |
| CN103810192A (en) * | 2012-11-09 | 2014-05-21 | 腾讯科技（深圳）有限公司 | User interest recommending method and device |
| CN105045931A (en) * | 2015-09-02 | 2015-11-11 | 南京邮电大学 | Video recommendation method and system based on Web mining |
| CN105512326A (en) * | 2015-12-23 | 2016-04-20 | 成都品果科技有限公司 | Picture recommending method and system |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107679883A (en) * | 2017-05-05 | 2018-02-09 | 平安科技（深圳）有限公司 | The method and system of advertisement generation |
| CN107688600A (en) * | 2017-07-12 | 2018-02-13 | 百度在线网络技术（北京）有限公司 | Knowledge point method for digging and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11030238B2 (en) | Determining and utilizing contextual meaning of digital standardized image characters | |
| Klostermann et al. | Extracting brand information from social networks: Integrating image, text, and social tagging data | |
| Culotta et al. | Mining brand perceptions from twitter social networks | |
| Chang et al. | Location3: How users share and respond to location-based data on social | |
| CN106886518B (en) | Microblog account number classification method | |
| US11514063B2 (en) | Method and apparatus of recommending information based on fused relationship network, and device and medium | |
| Mac Kim et al. | Demographic inference on twitter using recursive neural networks | |
| TWI724337B (en) | Computerized system and method for controlling electronic messages and their responses after delivery | |
| CN104717124A (en) | Friend recommendation method, device and server | |
| US10193849B2 (en) | Determining stories of interest based on quality of unconnected content | |
| CN105447193A (en) | Music recommending system based on machine learning and collaborative filtering | |
| US20150134663A1 (en) | Method, apparatus, and computer-readable storage medium for grouping social network nodes | |
| Zhang | Language in our time: An empirical analysis of hashtags | |
| US9900278B2 (en) | Eliciting positive responses to a social media posting | |
| CN105893484A (en) | Microblog Spammer recognition method based on text characteristics and behavior characteristics | |
| CN109344887A (en) | Short video classification methods, system and medium based on multi-modal dictionary learning | |
| Nitzsche et al. | Development of an Evaluation Tool for Participative E-Government Services: A Case Study of Electronic Participatory Budgeting Projects in Germany. | |
| Dong et al. | Discovering burst patterns of burst topic in twitter | |
| CN103473128A (en) | Collaborative filtering method for mashup application recommendation | |
| Ju et al. | Relationship strength estimation based on Wechat Friends Circle | |
| Vidal et al. | Application of social media for consumer research | |
| CN103279483B (en) | A kind of topic Epidemic Scope appraisal procedure towards micro-blog and system | |
| CN104899602A (en) | User cluster analysis system based on K-means algorithm | |
| CN106354858A (en) | Information resource recommendation method based on label clusters | |
| Yu et al. | Graph learning for fake review detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170125 |