CN103559269A - Knowledge recommending method for mobile news subscription - Google Patents
Knowledge recommending method for mobile news subscription Download PDFInfo
- Publication number
- CN103559269A CN103559269A CN201310538587.7A CN201310538587A CN103559269A CN 103559269 A CN103559269 A CN 103559269A CN 201310538587 A CN201310538587 A CN 201310538587A CN 103559269 A CN103559269 A CN 103559269A
- Authority
- CN
- China
- Prior art keywords
- knowledge
- knowledge point
- news
- module
- subscribe
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2455—Query execution
- G06F16/24552—Database cache management
Abstract
The invention relates to a knowledge recommending method for mobile news subscription. The method comprises the following steps: (1) sequentially extracting knowledge points from a knowledge base by an extraction module; (2) searching the names of the knowledge points from a common word dictionary by a search module, and generating a knowledge point set and storing the knowledge point set in a memory module by an integration module; (3) updating the classification information and the associated relation of the knowledge points of an identity in the memory module; (4) finding out all knowledge point names in a news document from the identity to form a set A by the search module; (5) disambiguating each element in the set A to form a set B by an analysis module; (6) filtering each element in the set B to form a recommendation list C by the analysis module; (7) outputting the recommendation list C by an output module. Through the utilization of the crowdsourcing knowledge base in the provision of the classification and relation information of the knowledge points, high timeliness, high usability and high expandability which are required by the background knowledge recommendation of mobile news subscription are effectively combined in a light-weight system.
Description
Technical field
The invention belongs to search field, specifically relate to a kind of knowledge recommend method towards mobile subscribe to news.
Background technology
The universal approach of greatly having expanded people's obtaining informations rapidly of the mobile intelligent terminal such as smart mobile phone and panel computer, particularly for news category information, the mode that people can subscribe to by keyword is obtained own interested content whenever and wherever possible.Yet user, when reading news, probably wishes the background knowledge (being generally the page of similar encyclopaedia entry character) of the related fact key element (as related person, organizational structure, place etc.) of this media event to do further understanding.At present, the technology that can meet this type of demand is mainly recommendation (or filtration) technology, and recommended technology is mainly divided into recommendation based on cooperation, content-based recommendation and the recommendation three types based on knowledge.
Recommendation based on cooperation: its basic assumption is that they also can have close preference in future if having close preference with some other user before user.The general arest neighbors that adopts (is pressed similarity degree-user based of user preference, then provide recommendation results according to the preference degree of prediction or similarity degree-item based of being had a preference for of article) or the preference of the method prediction active user of Mining Association Rules on new article (new knowledge point).But being most widely used of this type of recommendation, but need to obtain user preference data, the growth along with number of users and knowledge base scale, can sharply increase the demand of computational resource, and the content of knowledge point to be recommended is not considered in the recommendation based on cooperation.
Content-based recommendation: its basic assumption is that it may belong to the article that user likes if the article of certain article and user preference are close.This method is specially adapted to the recommendation of text.Conventionally adopt the vector space model based on tfidf to represent document, in application arest neighbors (or top k neighbour) method, provide recommendation list.But this type is recommended to be mainly used in text and is recommended, and knowledge page often comprises the information of multiple format, even comprises the content of dynamic change, is therefore difficult to application or builds a unified document (knowledge point) method for expressing (as tfidf).
Recommendation based on knowledge: the recommendation of this type is by its recommended requirements of user's explicit definition (forming the constraint to commodity), the similarity between system-computed user's request and article or adopt special rule to recommend.But the recommendation of this type is mainly for the commodity (as automobile, large electric appliances etc.) that can frequently do not bought in a period of time, because evaluation information seldom and easily lost efficacy in this class situation.The input of mobile phone users and selection are all difficulty compared with pc, allowing user oneself set complicated recommended requirements can increase the notice of the complexity of user's operation, long-time transferring user, simultaneously user to correlated knowledge point want ask far away from first purchase large scale commercial product want ask, therefore the recommended technology based on knowledge is unsuitable for the recommendation to knowledge point.
Summary of the invention
For the deficiencies in the prior art, the invention provides a kind of knowledge recommend method towards mobile subscribe to news.The features such as the fast response time, the visit capacity that require for mobile subscribe to news itself are large, content update is rapid, designed the knowledge recommend method based on mass-rent knowledge base and body, use the method can realize lightweight, has the backgrounding content recommendation system of high real-time, enhanced scalability, high availability.
The object of the invention is to adopt following technical proposals to realize:
A knowledge recommend method for mobile subscribe to news, its improvements are, described method comprises:
(1) extraction module extracts successively knowledge point from knowledge base;
(2) search module is from the title of everyday words dictionary lookup knowledge point and adopt integration module to generate knowledge point set to be stored in memory module;
(3) upgrade knowledge point classified information and the incidence relation of body in memory module;
(4) search module and from body, find out all knowledge points title formation set A in news documents;
(5) each the element disambiguation in analysis module pair set A forms set B;
(6) each element in analysis module pair set B filters and forms recommendation list C;
(7) output module output recommendation list C.
Preferably, described step (2) comprises by searching the title of knowledge point in everyday words dictionary, if everyday words continues to get next knowledge point; If it is not everyday words, is adopted integration module to generate knowledge point set and be stored in memory module, for recommendation.
Preferably, a knowledge point title may be to there being the polysemy of a plurality of knowledge points, and the mapping relations between knowledge point title and knowledge point id are safeguarded by memory module.
Preferably, described step (3) comprises according to body in knowledge point set, classificating word dictionary and three dictionaries generations of relative dictionary and renewal classification tree and knowledge point graph of a relation renewal memory module.
Further, described body comprises classification tree, graph of a relation, classificating word dictionary and relative dictionary; Knowledge point ID is positioned at the leaf node of ontology classification tree, and the intermediate node of classification tree is classifier ID, and classificating word dictionary comprises the mapping relations between classifier and its ID; Graph of a relation comprises target sparse matrix under the ID ranks of knowledge point, and matrix element is the list of relative ID, and relative dictionary comprises the mapping relations between relative and its ID.
Preferably, described step (5) comprises
If a, without ambiguity, adds candidate to recommend knowledge point set B corresponding knowledge point; With
If b ambiguity is selected to subscribe to user that knowledge point that knowledge point has minimum public ancestors from all knowledge points of correspondence, added set B.
Preferably, described step (6) comprises
If a) title occurrence number in news documents is more than or equal to threshold value δ, is added and recommended in knowledge point list C;
B) if title occurrence number is less than threshold value, but this knowledge point and user subscribe to knowledge point have correlationship (, added recommendation list C; With
C) if title occurrence number is less than threshold value, and subscribe to knowledge point without correlationship with user, abandon this knowledge point.
Further, described news documents is shorter, and for reducing empty recommendation, threshold value δ gets 0.
Compared with the prior art, beneficial effect of the present invention is:
1) algorithm of the present invention is simple, is easy to realize, and efficiency is high, is suitable for the environment higher to requirement of real-time such as mobile subscribe to news.
2) system magnitude of the present invention is light, without the user preference data of processing magnanimity, takies calculating and storage resources less.
3) extensibility of the present invention is strong, and main back-end data (everyday words dictionary, body, knowledge point set) scale is all little, and without the consistance that guarantees data
4) availability of the present invention is high, and everyday words dictionary, body, knowledge point set etc. all can online updatings.
5) knowledge base that the present invention builds body, based on mass-rent, had both guaranteed higher quality, again without using complicated natural language understanding technology.
6) the present invention takes full advantage of the classification of high-quality knowledge point and the relation information that mass-rent knowledge base provides, avoided obtaining, process and analyzing of large-scale consumer preference information, also avoided building the complicated natural language processing technique that Opening field body is commonly used, effectively recommended desired high real-time, high availability, enhanced scalability and high availability to be melted in the system of a lightweight background knowledge of mobile subscribe to news.
Accompanying drawing explanation
Fig. 1 is a kind of knowledge recommend method structured flowchart towards mobile subscribe to news provided by the invention.
Fig. 2 is a kind of process flow diagram towards knowledge processing in the knowledge recommend method of mobile subscribe to news provided by the invention.
Fig. 3 is provided by the invention a kind of towards recommending the process flow diagram of part in the knowledge recommend method of mobile subscribe to news.
Embodiment
Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described in further detail.
As shown in Figure 1, a kind of knowledge recommend method towards mobile subscribe to news of the present invention, for building body and knowledge point dictionary on the knowledge base basis based on mass-rent, carries out knowledge recommendation with it.System comprise two parts of knowledge processing and recommendation (comprise again content analysis and disambiguation/filter two main modular).
Knowledge processing partly utilizes everyday words dictionary (can manual maintenance) from knowledge base, to extract useful knowledge point (knowledge point of mating with everyday words is used as noise), and builds and upgrade body according to the classification in knowledge base and relation information;
Recommend part for the treatment of request (from picking out suitable recommendation word the document presenting for user), it finds out all available knowledge points as candidate's recommended project in news documents, then utilize the constructed body of knowledge processing and filtering rule, candidate's recommended project is carried out to disambiguation and filtration, to improve the precision of recommendation.
Body comprises classification tree, graph of a relation, classificating word dictionary and relative dictionary.Knowledge point id is positioned at the leaf node of ontology classification tree, and the intermediate node of classification tree is classifier id, and the mapping relations between classifier and its id are deposited by classificating word dictionary; Graph of a relation can be stored by make target sparse matrix under ranks with knowledge point id, and matrix element is the list of relative id, and the mapping relations between relative and its id are deposited by relative dictionary.
Due to the existence of polysemy, a knowledge point title may be to there being a plurality of knowledge points (they have different knowledge point id), and the mapping relations between knowledge point title and knowledge point id can be safeguarded by knowledge point set.
As shown in Figure 2, the flow process key step of knowledge processing comprises:
1) from knowledge base, take out successively each knowledge point;
2) in everyday words dictionary, search the title of this knowledge point, if it is everyday words, continue to get next knowledge point.If it is not everyday words, added knowledge point set, for recommendation;
3) to adding the knowledge point of knowledge point set, get its classified information and incidence relation, upgrade body (generating and renewal classification tree and knowledge point graph of a relation by knowledge point set, classificating word dictionary and three dictionaries of relative dictionary).
As shown in Figure 3, the flow process key step of recommendation comprises as follows:
1) utilize knowledge point set, find out all knowledge points title in news documents, in income set A.
2) each element in pair set A (knowledge point title) carries out disambiguation operation;
If a) it,, without ambiguity (a corresponding knowledge point id), adds candidate to recommend knowledge point set B its corresponding knowledge point.
B) if ambiguity (corresponding a plurality of knowledge point id) (is utilized the classification tree of body) and selected to subscribe to user that knowledge point that knowledge point has minimum public ancestors from all knowledge points of its correspondence, added set B.
3) each element (knowledge point) in pair set B carries out filter operation;
If a) its title occurrence number in news documents is more than or equal to threshold value δ (as getting 3), added (for newsflash document, for reducing empty recommendation, threshold value δ gets 0) in the recommendation list C of knowledge point;
B) if its title occurrence number is less than threshold value, but subscribing to knowledge point, this knowledge point and user have correlationship (determining with the graph of a relation of body), added recommendation list C;
C) if title occurrence number is less than threshold value, and subscribe to knowledge point without correlationship with user, abandon this knowledge point.
4) output recommendation list C.
Embodiment
User a is interested in described certain the class news of keyword w, and he (she) can only subscribe to this keyword in the subscribe to news software S of mobile device (as smart mobile phone) at it.The basic function of S is that the up-to-date news t relevant to w presented to a.Existing S wishes to increase background knowledge recommendation function, when t is presented to a, the relevant structural knowledge of some the word/phrase p in t (may comprise the basic introduction of text formatting, to the hyperlink of other related pages and the link of various content of multimedia etc.) is also recommended to a.
Implementation method: utilize the set P of mass-rent mode construction knowledge base K(p, the related content that the page of each p contains classified information, various forms and with the correlationship of other p-as related person etc., and the page elements of each p may be different with layout).Utilize artificial or semi-artificial mode to build general term dictionary D.Adopt the knowledge process flow (utilizing D) shown in Fig. 2 from K, to construct knowledge point set Z and body O.For news t, before presenting to a, can use the recommended flowsheet shown in Fig. 3 (utilizing Z and D) to provide and recommend knowledge point list, present to afterwards a together with t.Concrete presentation mode can adopt in t the appearance for the first time of word in recommendation list is presented as anchor text, also can t present region near (as below) list and recommend the mode of word to present.
Finally should be noted that: above embodiment is only in order to illustrate that technical scheme of the present invention is not intended to limit, although the present invention is had been described in detail with reference to above-described embodiment, those of ordinary skill in the field are to be understood that: still can modify or be equal to replacement the specific embodiment of the present invention, and do not depart from any modification of spirit and scope of the invention or be equal to replacement, it all should be encompassed in the middle of claim scope of the present invention.
Claims (8)
1. towards a knowledge recommend method for mobile subscribe to news, it is characterized in that, described method comprises:
(1) extraction module extracts successively knowledge point from knowledge base;
(2) search module is from the title of everyday words dictionary lookup knowledge point and adopt integration module to generate knowledge point set to be stored in memory module;
(3) upgrade knowledge point classified information and the incidence relation of body in memory module;
(4) search module and from body, find out all knowledge points title formation set A in news documents;
(5) each the element disambiguation in analysis module pair set A forms set B;
(6) each element in analysis module pair set B filters and forms recommendation list C;
(7) output module output recommendation list C.
2. a kind of knowledge recommend method towards mobile subscribe to news as claimed in claim 1, is characterized in that, described step (2) comprises by searching the title of knowledge point in everyday words dictionary, if everyday words continues to get next knowledge point; If it is not everyday words, is adopted integration module to generate knowledge point set and be stored in memory module, for recommendation.
3. a kind of knowledge recommend method towards mobile subscribe to news as claimed in claim 1, it is characterized in that, a knowledge point title may be to there being the polysemy of a plurality of knowledge points, and the mapping relations between knowledge point title and knowledge point id are safeguarded by memory module.
4. a kind of knowledge recommend method towards mobile subscribe to news as claimed in claim 1, it is characterized in that, described step (3) comprises according to body in knowledge point set, classificating word dictionary and three dictionaries generations of relative dictionary and renewal classification tree and knowledge point graph of a relation renewal memory module.
5. a kind of knowledge recommend method towards mobile subscribe to news as claimed in claim 4, is characterized in that, described body comprises classification tree, graph of a relation, classificating word dictionary and relative dictionary; Knowledge point ID is positioned at the leaf node of ontology classification tree, and the intermediate node of classification tree is classifier ID, and classificating word dictionary comprises the mapping relations between classifier and its ID; Graph of a relation comprises target sparse matrix under the ID ranks of knowledge point, and matrix element is the list of relative ID, and relative dictionary comprises the mapping relations between relative and its ID.
6. a kind of knowledge recommend method towards mobile subscribe to news as claimed in claim 1, is characterized in that, described step (5) comprises
If a, without ambiguity, adds candidate to recommend knowledge point set B corresponding knowledge point; With
If b ambiguity is selected to subscribe to user that knowledge point that knowledge point has minimum public ancestors from all knowledge points of correspondence, added set B.
7. a kind of knowledge recommend method towards mobile subscribe to news as claimed in claim 1, is characterized in that, described step (6) comprises
If a) title occurrence number in news documents is more than or equal to threshold value δ, is added and recommended in knowledge point list C;
B) if title occurrence number is less than threshold value, but this knowledge point and user subscribe to knowledge point have correlationship (, added recommendation list C; With
C) if title occurrence number is less than threshold value, and subscribe to knowledge point without correlationship with user, abandon this knowledge point.
8. a kind of knowledge recommend method towards mobile subscribe to news as claimed in claim 7, is characterized in that, described news documents is shorter, and for reducing empty recommendation, threshold value δ gets 0.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310538587.7A CN103559269B (en) | 2013-11-04 | 2013-11-04 | A kind of knowledge recommendation method towards mobile news subscription |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310538587.7A CN103559269B (en) | 2013-11-04 | 2013-11-04 | A kind of knowledge recommendation method towards mobile news subscription |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103559269A true CN103559269A (en) | 2014-02-05 |
| CN103559269B CN103559269B (en) | 2018-02-06 |
Family
ID=50013515
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310538587.7A Expired - Fee Related CN103559269B (en) | 2013-11-04 | 2013-11-04 | A kind of knowledge recommendation method towards mobile news subscription |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103559269B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105302897A (en) * | 2015-10-21 | 2016-02-03 | 无锡天脉聚源传媒科技有限公司 | Search result acquisition method and apparatus |
| CN107967254A (en) * | 2017-10-31 | 2018-04-27 | 科大讯飞股份有限公司 | Knowledge point Forecasting Methodology and device, storage medium, electronic equipment |
| CN108536872A (en) * | 2018-04-27 | 2018-09-14 | 武汉文都信息技术有限公司 | Optimize the method and apparatus of knowledge base structure |
| CN111737407A (en) * | 2020-08-25 | 2020-10-02 | 成都数联铭品科技有限公司 | Event unique ID construction method based on event disambiguation |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1998020405A2 (en) * | 1996-11-06 | 1998-05-14 | U.C. Media, Inc. | Method and system for screen text capturing |
| CN102611785A (en) * | 2011-01-20 | 2012-07-25 | 北京邮电大学 | Personalized active news recommending service system and method for mobile phone user |
-
2013
- 2013-11-04 CN CN201310538587.7A patent/CN103559269B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1998020405A2 (en) * | 1996-11-06 | 1998-05-14 | U.C. Media, Inc. | Method and system for screen text capturing |
| CN102611785A (en) * | 2011-01-20 | 2012-07-25 | 北京邮电大学 | Personalized active news recommending service system and method for mobile phone user |
Non-Patent Citations (4)
| Title |
|---|
| 刘振: "社交媒体的语义标注应用现状研究", 《图书与情报》 * |
| 周康 等: "基于移动互联网的文献个性化推荐系统", 《计算机应用》 * |
| 戚欣 等: "基于本体知识库的自动语义标注", 《计算机应用研究》 * |
| 艾丹祥: "基于本体论的知识检索研究", 《中国优秀博硕士学位论文全文数据库（硕士） 信息科技辑》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105302897A (en) * | 2015-10-21 | 2016-02-03 | 无锡天脉聚源传媒科技有限公司 | Search result acquisition method and apparatus |
| CN105302897B (en) * | 2015-10-21 | 2018-11-20 | 无锡天脉聚源传媒科技有限公司 | A kind of acquisition methods and device of search result |
| CN107967254A (en) * | 2017-10-31 | 2018-04-27 | 科大讯飞股份有限公司 | Knowledge point Forecasting Methodology and device, storage medium, electronic equipment |
| CN107967254B (en) * | 2017-10-31 | 2021-05-04 | 科大讯飞股份有限公司 | Knowledge point prediction method and device, storage medium and electronic equipment |
| CN108536872A (en) * | 2018-04-27 | 2018-09-14 | 武汉文都信息技术有限公司 | Optimize the method and apparatus of knowledge base structure |
| CN108536872B (en) * | 2018-04-27 | 2020-12-25 | 湖北时代万新国际教育发展有限公司 | Method and device for optimizing knowledge base structure |
| CN111737407A (en) * | 2020-08-25 | 2020-10-02 | 成都数联铭品科技有限公司 | Event unique ID construction method based on event disambiguation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103559269B (en) | 2018-02-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7201730B2 (en) | Intention recommendation method, device, equipment and storage medium | |
| CN110321482B (en) | Information recommendation method, device and equipment | |
| US20220261427A1 (en) | Methods and system for semantic search in large databases | |
| KR101166130B1 (en) | Method and apparatus for constructing user profile using contents tag, and method for contents recommendation using the constructed user profile | |
| CN101271459B (en) | Word library generation method, input method and input method system | |
| Shi et al. | Learning-to-rank for real-time high-precision hashtag recommendation for streaming news | |
| US10482146B2 (en) | Systems and methods for automatic customization of content filtering | |
| US20140379719A1 (en) | System and method for tagging and searching documents | |
| CN112052387B (en) | Content recommendation method, device and computer readable storage medium | |
| CN103123649A (en) | Method and system for searching information based on micro blog platform | |
| CN108090178B (en) | Text data analysis method, text data analysis device, server and storage medium | |
| WO2014206151A1 (en) | System and method for tagging and searching documents | |
| CN103455487A (en) | Extracting method and device for search term | |
| CN110232126B (en) | Hot spot mining method, server and computer readable storage medium | |
| CN103384883A (en) | Semantic enrichment by exploiting Top-K processing | |
| Martín et al. | Using semi-structured data for assessing research paper similarity | |
| US20230386238A1 (en) | Data processing method and apparatus, computer device, and storage medium | |
| CN109582869A (en) | A kind of data processing method, device and the device for data processing | |
| CN112069326A (en) | Knowledge graph construction method and device, electronic equipment and storage medium | |
| CN103559269A (en) | Knowledge recommending method for mobile news subscription | |
| Celikyilmaz et al. | Leveraging web query logs to learn user intent via bayesian latent variable model | |
| Pannu et al. | Explicit user profiles in web search personalisation | |
| KR101651963B1 (en) | Method of generating time and space associated data, time and space associated data generation server performing the same and storage medium storing the same | |
| Liu et al. | Recent advances in personal recommender systems | |
| CN109977294B (en) | Information/query processing device, query processing/text query method, and storage medium |
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 | ||
| C41 | Transfer of patent application or patent right or utility model | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20160412 Address after: 100086, No. 2, building 43, No. 5 West Third Ring Road, Haidian District, Beijing, 01-03A Applicant after: Beijing Wyatt Network Technology Co.,Ltd. Address before: Shou Heng Technology Building No. 51 Beijing 100191 Haidian District Xueyuan Road room 0902 Applicant before: BEIJING ZHONGSOU NETWORK TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180206 Termination date: 20211104 |