CN106294868A - A kind of personalized recommendation method based on search engine and system - Google Patents
A kind of personalized recommendation method based on search engine and system Download PDFInfo
- Publication number
- CN106294868A CN106294868A CN201610712505.XA CN201610712505A CN106294868A CN 106294868 A CN106294868 A CN 106294868A CN 201610712505 A CN201610712505 A CN 201610712505A CN 106294868 A CN106294868 A CN 106294868A
- Authority
- CN
- China
- Prior art keywords
- query word
- word
- search query
- result
- module
- 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
- G06F16/957—Browsing optimisation, e.g. caching or content distillation
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of personalized recommendation method based on search engine and system, method comprises the steps: to obtain active user by the query word searched for by search engine;Search query word is carried out natural language processing, obtains key word fragment and corresponding weight;By by the off-line Result on backstage, click feedback data and described key word fragment, after being weighted and sorting, export recommendation results.The present invention can be different for the search query word of different users calculated associated recommendation result, therefore is applicable to the associated recommendation in multiple field, such as video, news, advertisement etc..By by the off-line Result on backstage, click feedback data and described key word fragment, after being weighted and sorting, associated recommendation result can be changed in real time along with the search inquiry of user, i.e. can be instantly obtained different result when the search query word that different user request is different, bring good Consumer's Experience to user.
Description
Technical field
The present invention relates to commending contents field, particularly to personalized recommendation method based on search engine and system.
Background technology
It is presently recommended that system is the most all by backstage calculated off line, then the recommendation results excavated in advance is showed
User, but how can quickly respond the demand of user, and real-time adjusts the recommendation results for different user, becomes existing
In commending system problem encountered and challenge.
Summary of the invention
The technical problem to be solved in the present invention is, for the request with search engine inquiry word, according to corresponding search
Query word calculates in real time and adjusts, and finally gives the real-time recommendation result of related content.
Solve above-mentioned technical problem, the invention provides a kind of personalized recommendation method based on search engine, including such as
Lower step:
Obtain active user by the query word searched for by search engine;
Search query word is carried out natural language processing, obtains key word fragment and corresponding weight;
By by the off-line Result on backstage, click feedback data and described key word fragment, being weighted and sort
After, export recommendation results.
Further, described active user refers to listed user.
Further, described click feedback data, in order to according to recommendation results before, to obtain user click condition score,
If clicked number of times score at most is high, and the result preferential recommendation high to score.
Further, described off-line Result, in order to give a mark according to the total quality of recommended content, described
Total quality includes: the abundant degree of content and high-quality degree.
Based on above-mentioned, the invention provides a kind of personalized recommendation system based on search engine, including: search query word
Module, search query word parsing module, real-time online computing module,
Described search query word module, in order to obtain active user by the query word searched for by search engine;
Described search query word parsing module, in order to search query word is carried out natural language processing, obtains key word broken
Sheet and corresponding weight;
Described real-time online computing module, in order to by by the off-line Result on backstage, click on feedback data and described
Key word fragment, after being weighted and sorting, exports recommendation results.
Further, system also includes off-line Result module, in order to enter according to the total quality of recommended content
Row marking, and result is inputted real-time online computing module.
Further, system also include click on feedback data module, in order to according to mouse click event obtain user for
The Real-time Feedback result of the search query word in described search query word module.
Further, described search query word parsing module also in order to, search query word is carried out natural language processing, logical
Crossing Chinese word segmentation, TF/IDF, one or more methods in proper name dictionary calculate and extract key word fragment.
Further, described real-time online computing module, also in order to by merger and operation of reordering, export associated recommendation
Result.
Further, described search query word module is connected with search engine API, in order to obtain query word
Beneficial effects of the present invention:
1) due to a kind of based on search engine the personalized recommendation method in the present invention, pass through owing to obtaining active user
The query word searched for by search engine;Search query word is carried out natural language processing, obtains key word fragment and relatively
The weight answered;By by the off-line Result on backstage, click feedback data and described key word fragment, being weighted and sort
After, export recommendation results.By different search query word, thus it is possible to vary the associated recommendation result of user, for different use
The search query word calculated associated recommendation result at family is different, therefore is applicable to the associated recommendation in multiple field, such as
Video, news, advertisement etc..By by the off-line Result on backstage, click feedback data and described key word fragment, adding
After weighing and sorting so that associated recommendation result can change in real time along with the search inquiry of user, i.e. ask when different user
Different result can be instantly obtained during different search query word, bring good Consumer's Experience to user.
2) personalized recommendation system based on search engine in the present invention, owing to including search query word module, search
Query word parsing module, real-time online computing module, by described real-time online computing module, in order to by by the off-line on backstage
Result, click feedback data and described key word fragment so that associated recommendation result can be real along with the search inquiry of user
Time change, i.e. can be instantly obtained different result when the different user different search query word of request, carry to user
Carry out good Consumer's Experience.By described search query word module, obtain active user by looking into of being searched for by search engine
Ask word;By described search query word parsing module, search query word is carried out natural language processing, obtain key word fragment with
And corresponding weight, and result is input in real-time online computing module by different search query word, thus it is possible to vary
The associated recommendation result of user, thus the search query word calculated associated recommendation result for different users is different
's.
Accompanying drawing explanation
Fig. 1 is personalized recommendation method schematic flow sheet based on search engine in one embodiment of the invention.
Fig. 2 is personalized recommendation system structural representation based on search engine in one embodiment of the invention.
Fig. 3 is personalized recommendation system structural representation based on search engine in another embodiment of the present invention.
Detailed description of the invention
For making the object, technical solutions and advantages of the present invention clearer, below in conjunction with specific embodiment, and reference
Accompanying drawing, the present invention is described in more detail.
Fig. 1 is personalized recommendation method schematic flow sheet based on search engine in one embodiment of the invention.
A kind of based on search engine personalized recommendation method in the present embodiment, comprises the steps:
Step S100 obtains active user by the query word searched for by search engine;Element engine type is searched with concrete
Unrelated, user has searched for query word in a search engine, and after then clicking in the page of corresponding website, website i.e. can be known
Road is from which query word.Those skilled in the art can understand, detailed process is i.e. that after user searches query word, search engine makes
With term, inside the server of oneself, find out the webpage mated with this index terms.Preferably, first by data set according to
Hash method resolves into multiple small data set, then uses trie tree or hash to add up the query word in each small data set
Frequently, obtain with little top heap afterwards and each data set goes out the front K number that frequency is the highest, finally obtain final in all top K
Top K, obtain query word.In certain embodiments, described active user refers to listed user, it is simple to obtain user's
Query word statistics and custom.
Step S101 carries out natural language processing to search query word, obtains key word fragment and corresponding weight;
Those skilled in the art can understand, the natural language processing carrying out described query word includes but not limited to: divided by Chinese
Word, TF/IDF, one or more in proper name dictionary, query word is carried out cutting and obtains key word fragment.Described Chinese divides
Word (Chinese Word Segmentation) refers to be cut into a Chinese character sequence one by one individually word.Chinese
Participle is exactly the process that continuous print word sequence is reassembled into word sequence according to certain specification.Chinese word segmentation draws for search
For holding up, most important is not to find all results, because it is too many to find all results not have in the webpage of over ten billion
Meaning, nobody can see complete, it is most important that maximally related result is come foremost, and this is also referred to as relevancy ranking.In
Literary composition participle whether accurate, usually directly influence the relevancy ranking to Search Results.For qualitative analysis, search engine
Segmentation methods different, the difference of dictionary all can affect the return result of the page.Described TF/IDF, TF represent word frequency.Weight IDF
Being called " inverse document frequency " (Inverse Document Frequency), its size becomes anti-with the common degree of a word
Ratio.
The first step, calculates word frequency.
Word frequency (TF)=certain word occurrence number in article
There is dividing of length in view of article, for the ease of the comparison of different articles, carry out " word frequency " standardization.
The total degree of word frequency (TF)=certain word occurrence number/article in article
Or
The occurrence number of the word that the number of times that word frequency (TF)=certain word occurs in article/this article occurrence number is most
Second step, calculates inverse document frequency.
Inverse document frequency (IDF)=log (number of files+1 of this word of the total number of documents of corpus/comprise)
3rd step, calculates TF-IDF.
TF-ID=word frequency (TF) * inverse document frequency (IDF)
It will be seen that the occurrence number that TF-IDF and word are in a document is directly proportional, with this word in whole language
Occurrence number is inversely proportional to.So, the algorithm automatically extracting key word just will be apparent from, it is simply that calculates each word of document
TF-IDF value, arranges the most in descending order, takes the several words coming foremost.
Described proper name dictionary, refers to proper noun dictionary, in order to query word is carried out fragmentation more accurately.
Step S102 is by by the off-line Result on backstage, click feedback data and described key word fragment, adding
After weighing and sorting, export recommendation results.In certain embodiments, described click feedback data, in order to basis recommendation results before,
Obtain user click condition score, if clicked number of times score at most is high, and the result preferential recommendation high to score.Real at some
Execute in example, described off-line Result, in order to give a mark according to the total quality of recommended content, described total quality bag
Include: the abundant degree of content and high-quality degree.
Based on features described above, the technique effect in the present embodiment at least includes: by different search query word, Ke Yigai
Becoming the associated recommendation result of user, the search query word calculated associated recommendation result for different users is different,
Therefore it is applicable to the associated recommendation in multiple field, such as video, news, advertisement etc..By by the off-line Result on backstage, point
Hit feedback data and described key word fragment, after being weighted and sorting so that associated recommendation result can be along with the search of user
Inquiry changes in real time, i.e. can be instantly obtained different result when the search query word that different user request is different, give
User brings good Consumer's Experience.
Fig. 2 is personalized recommendation system structural representation based on search engine in one embodiment of the invention.
A kind of based on search engine personalized recommendation system in the present embodiment, including: search query word module 1, searches
Rope query word parsing module 2, real-time online computing module 3,
Described search query word module 1, in order to obtain active user by the query word searched for by search engine;
Described search query word parsing module 2, in order to search query word is carried out natural language processing, obtains key word broken
Sheet and corresponding weight;Preferred as in the present embodiment, described search query word parsing module 2 also in order to, to search
Query word carries out natural language processing, and by Chinese word segmentation, TF/IDF, one or more methods in proper name dictionary calculate and take out
Take out key word fragment.Preferred as in the present embodiment, described search query word module 2 is connected with search engine API, in order to
Obtain query word.
Described real-time online computing module 3, in order to by by the off-line Result on backstage, click on feedback data and described
Key word fragment, after being weighted and sorting, exports recommendation results.Preferred as in the present embodiment, described real-time online meter
Calculate module 3, also in order to by merger and operation of reordering, export associated recommendation result.
Preferred as in the present embodiment, system also includes off-line Result module 4, in order to according to recommended content
Total quality give a mark, and result is inputted real-time online computing module.
Preferred as in the present embodiment, system also includes clicking on feedback data module 5, in order to according to mouse click event
Obtain user's Real-time Feedback result for the search query word in described search query word module.
Specifically,
Search query word module 1 mainly receives the search query word of user.
Such as: user Steve,
One section of query word XXX is inputted by search engine.
Search query word parsing module 2 can carry out natural language processing to search query word.
By Chinese word segmentation, TF/IDF, proper name dictionary etc. calculates and extracts key word A, and B, C obtain Term Fragment
Multiple key word fragments and weight
Many key word fragments of Term Fragment and weight are as follows:
Key word A, weight 0.5
Key word B, weight 0.3
Key word C, weight 0.2
Real-time online computing module 3 obtains each key word and the weight of active user's search query word, in conjunction with offline
Result off-line Result and click feedback click on feedback data, then they are weighted ordering by merging,
Output recommendation results eventually.
Such as: user Steve,
Key word A, weight 0.5, obtain correlated results:
B=0.31, g=0.19
Key word B, weight 0.3, obtain correlated results:
X=0.16, y=0.14
Key word C, weight 0.2, obtain correlated results:
T=0.18, u=0.02
Offline result off-line Result:
X=0.52, s=0.48, g=0.27, k=0.13, t=0.12
Click feedback clicks on feedback data:
X=0.21, u=0.10, p=0.07
Reordered operation by merge merger and rerank, finally for the associated recommendation result of Steve:
{ b=0.31, g=0.19}+{x=0.16, y=0.14}+{t=0.18, u=0.02}+{x=0.52, s=
0.48, g=0.27, k=0.14, t=0.12}+{x=0.21, u=0.11, p=0.07}
={ x=0.89, s=0.48, g=0.46, b=0.31, t=0.30, y=0.14, k=0.13, u=0.12, p
=0.07}
Those skilled in the art can understand, above-mentioned alphabetical a-z represents different Candidate Recommendation results respectively.
Personalized recommendation system based on search engine in the present embodiment, by described real-time online computing module, uses
With by by the off-line Result on backstage, click on feedback data and described key word fragment so that associated recommendation result can be with
The search inquiry user changes in real time, i.e. can be instantly obtained when the search query word that different user request is different and differ
The result of sample, brings good Consumer's Experience to user.By described search query word module, obtain active user and pass through
The query word of search engine search;By described search query word parsing module, search query word is carried out natural language processing,
Obtain key word fragment and corresponding weight, and result is input in real-time online computing module by different search
Query word, thus it is possible to vary the associated recommendation result of user, thus for the calculated phase of search query word of different users
It is different for closing recommendation results.
Fig. 3 is personalized recommendation system structural representation based on search engine in another embodiment of the present invention.
A kind of based on search engine personalized recommendation system in the present embodiment, including: search query word module 1, searches
Rope query word parsing module 2, real-time online computing module 3,
Described search query word module 1, in order to obtain active user by the query word searched for by search engine;Will knot
Fruit is input to described search query word parsing module 2, in order to search query word is carried out natural language processing, obtains key word broken
Sheet and corresponding weight;Obtain key word fragment 1, key word fragment 2, key word fragment 3;By key word fragment 1, key
Word fragment 2, key word fragment 3 input described real-time online computing module 3, in order to by by the off-line Result 4 on backstage, point
Hit feedback data 5 and described key word fragment, after being weighted and sorting, export recommendation results.
Those of ordinary skill in the field it is understood that more than, described be only the present invention specific embodiment, and
Be not used in the restriction present invention, all within the spirit and principles in the present invention, any modification, equivalent substitution and improvement etc. done, all
Within protection scope of the present invention should being included in.
Claims (10)
1. a personalized recommendation method based on search engine, it is characterised in that comprise the steps:
Obtain active user by the query word searched for by search engine;
Search query word is carried out natural language processing, obtains key word fragment and corresponding weight;
By by the off-line Result on backstage, click feedback data and described key word fragment, after being weighted and sorting, defeated
Go out recommendation results.
Personalized recommendation method the most according to claim 1, it is characterised in that described active user refers to listed use
Family.
Personalized recommendation method the most according to claim 1, it is characterised in that described click feedback data, in order to basis
Recommendation results before, obtains user click condition score, if clicked number of times score at most is high and preferential to the result that score is high
Recommend.
Personalized recommendation method the most according to claim 1, it is characterised in that described off-line Result, in order to basis
The total quality of recommended content is given a mark, and described total quality includes: the abundant degree of content and high-quality degree.
5. a personalized recommendation system based on search engine, it is characterised in that including: search query word module, search are looked into
Ask word parsing module, real-time online computing module,
Described search query word module, in order to obtain active user by the query word searched for by search engine;
Described search query word parsing module, in order to search query word is carried out natural language processing, obtain key word fragment with
And corresponding weight;
Described real-time online computing module, in order to by by the off-line Result on backstage, click feedback data and described key
Word fragment, after being weighted and sorting, exports recommendation results.
Personalized recommendation system the most according to claim 5, it is characterised in that also include off-line Result module, uses
Give a mark with the total quality according to recommended content, and result is inputted real-time online computing module.
Personalized recommendation system the most according to claim 5, it is characterised in that also include clicking on feedback data module, use
To obtain user's Real-time Feedback result for the search query word in described search query word module according to mouse click event.
Personalized recommendation method the most according to claim 1, it is characterised in that described search query word parsing module is also used
With, search query word is carried out natural language processing, by Chinese word segmentation, TF/IDF, one or more in proper name dictionary
Method calculates and extracts key word fragment.
Personalized recommendation method the most according to claim 1, it is characterised in that described real-time online computing module, also uses
With by merger and operation of reordering, export associated recommendation result.
Personalized recommendation method the most according to claim 1, it is characterised in that described search query word module and search
Engine API connects, in order to obtain query word.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610712505.XA CN106294868A (en) | 2016-08-23 | 2016-08-23 | A kind of personalized recommendation method based on search engine and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610712505.XA CN106294868A (en) | 2016-08-23 | 2016-08-23 | A kind of personalized recommendation method based on search engine and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106294868A true CN106294868A (en) | 2017-01-04 |
Family
ID=57615936
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610712505.XA Pending CN106294868A (en) | 2016-08-23 | 2016-08-23 | A kind of personalized recommendation method based on search engine and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106294868A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106886583A (en) * | 2017-02-14 | 2017-06-23 | 广州三星通信技术研究有限公司 | Method and apparatus for managing multimedia file broadcasting in the electronic device |
CN107194572A (en) * | 2017-05-18 | 2017-09-22 | 广东中科国志科技发展有限公司 | A kind of method and apparatus of technology evaluation |
CN108874888A (en) * | 2017-05-15 | 2018-11-23 | 李建文 | Data searching method |
CN111400589A (en) * | 2020-03-06 | 2020-07-10 | 孙月华 | Intelligent service platform for sports information |
CN112182403A (en) * | 2020-10-23 | 2021-01-05 | 网易(杭州)网络有限公司 | Question query method, device, equipment and storage medium |
US10956502B2 (en) | 2018-05-18 | 2021-03-23 | Yandex Europe Ag | Method of and system for recommending fresh search query suggestions on search engine |
CN112925967A (en) * | 2021-02-07 | 2021-06-08 | 北京鼎诚世通科技有限公司 | Method, device and equipment for generating expanded query words and storage medium |
CN115796172A (en) * | 2023-02-08 | 2023-03-14 | 广东粤港澳大湾区硬科技创新研究院 | Fault case recommendation method, device and system |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102685565A (en) * | 2012-05-18 | 2012-09-19 | 合一网络技术(北京)有限公司 | Click feedback type individual recommendation system |
CN102955857A (en) * | 2012-11-09 | 2013-03-06 | 北京航空航天大学 | Class center compression transformation-based text clustering method in search engine |
CN104216931A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津)科技有限公司 | Real-time recommending system and method |
CN105488221A (en) * | 2015-12-25 | 2016-04-13 | 北京奇虎科技有限公司 | Method and system for recommending query terms for conducting searching in search interface |
CN105574015A (en) * | 2014-10-13 | 2016-05-11 | 阿里巴巴集团控股有限公司 | Search recommendation method and device |
-
2016
- 2016-08-23 CN CN201610712505.XA patent/CN106294868A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102685565A (en) * | 2012-05-18 | 2012-09-19 | 合一网络技术(北京)有限公司 | Click feedback type individual recommendation system |
CN102955857A (en) * | 2012-11-09 | 2013-03-06 | 北京航空航天大学 | Class center compression transformation-based text clustering method in search engine |
CN104216931A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津)科技有限公司 | Real-time recommending system and method |
CN105574015A (en) * | 2014-10-13 | 2016-05-11 | 阿里巴巴集团控股有限公司 | Search recommendation method and device |
CN105488221A (en) * | 2015-12-25 | 2016-04-13 | 北京奇虎科技有限公司 | Method and system for recommending query terms for conducting searching in search interface |
Non-Patent Citations (2)
Title |
---|
李小平等: "《网络影视课程编导论》", 30 April 2016, 北京理工大学出版社 * |
陈洪辉等: "《指挥控制信息精准服务》", 30 June 2015, 国防工业出版社 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106886583A (en) * | 2017-02-14 | 2017-06-23 | 广州三星通信技术研究有限公司 | Method and apparatus for managing multimedia file broadcasting in the electronic device |
CN106886583B (en) * | 2017-02-14 | 2019-11-26 | 广州三星通信技术研究有限公司 | Method and apparatus for managing multimedia file broadcasting in the electronic device |
CN108874888A (en) * | 2017-05-15 | 2018-11-23 | 李建文 | Data searching method |
CN107194572A (en) * | 2017-05-18 | 2017-09-22 | 广东中科国志科技发展有限公司 | A kind of method and apparatus of technology evaluation |
US10956502B2 (en) | 2018-05-18 | 2021-03-23 | Yandex Europe Ag | Method of and system for recommending fresh search query suggestions on search engine |
CN111400589A (en) * | 2020-03-06 | 2020-07-10 | 孙月华 | Intelligent service platform for sports information |
CN112182403A (en) * | 2020-10-23 | 2021-01-05 | 网易(杭州)网络有限公司 | Question query method, device, equipment and storage medium |
CN112182403B (en) * | 2020-10-23 | 2023-09-15 | 网易(杭州)网络有限公司 | Question query method, device, equipment and storage medium |
CN112925967A (en) * | 2021-02-07 | 2021-06-08 | 北京鼎诚世通科技有限公司 | Method, device and equipment for generating expanded query words and storage medium |
CN115796172A (en) * | 2023-02-08 | 2023-03-14 | 广东粤港澳大湾区硬科技创新研究院 | Fault case recommendation method, device and system |
CN115796172B (en) * | 2023-02-08 | 2023-06-02 | 广东粤港澳大湾区硬科技创新研究院 | Fault case recommendation method, device and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106294868A (en) | A kind of personalized recommendation method based on search engine and system | |
CN101593200B (en) | Method for classifying Chinese webpages based on keyword frequency analysis | |
CN103678576B (en) | The text retrieval system analyzed based on dynamic semantics | |
US8346792B1 (en) | Query generation using structural similarity between documents | |
TWI508011B (en) | Category information providing method and device | |
CN103257957B (en) | A kind of text similarity recognition methods and device based on Chinese word segmentation | |
CN102043845B (en) | Method and equipment for extracting core keywords based on query sequence cluster | |
CN101639857B (en) | Method, device and system for establishing knowledge questioning and answering sharing platform | |
CN104199833B (en) | The clustering method and clustering apparatus of a kind of network search words | |
CN101986306B (en) | Method and equipment for acquiring yellow page information based on query sequence | |
CN104216931A (en) | Real-time recommending system and method | |
JP2006004417A (en) | Method and device for recognizing specific type of information file | |
JP2013529805A5 (en) | Search method, search system and computer program | |
CN103838756A (en) | Method and device for determining pushed information | |
EP2382534A1 (en) | Search engine for refining context-based queries based upon historical user feedback | |
WO2017157200A1 (en) | Characteristic keyword extraction method and device | |
CN107330057B (en) | Optimization method and system for ElasticSearch search correlation algorithm | |
CN104268175A (en) | Data search device and method thereof | |
CN106959949B (en) | A kind of data structured processing method for recommender system | |
CN103377185A (en) | Method and device for adding tags to short texts automatically | |
CN107844580A (en) | A kind of search term matching process | |
CN108388556B (en) | Method and system for mining homogeneous entity | |
CN102722502A (en) | Method and device for obtaining redundance-free suggested inquiry sequence | |
JP6049201B2 (en) | Program, method, apparatus and server for generating co-occurrence patterns for detecting synonyms | |
CN106951511A (en) | A kind of Text Clustering Method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170104 |