CN103823893A - User comment-based product search method and system - Google Patents

User comment-based product search method and system Download PDF

Info

Publication number
CN103823893A
CN103823893A CN201410086745.4A CN201410086745A CN103823893A CN 103823893 A CN103823893 A CN 103823893A CN 201410086745 A CN201410086745 A CN 201410086745A CN 103823893 A CN103823893 A CN 103823893A
Authority
CN
China
Prior art keywords
product
comment
user
data
score value
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
Application number
CN201410086745.4A
Other languages
Chinese (zh)
Inventor
闫宏飞
赵鑫
江翰
李晓明
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Peking University
Original Assignee
Peking University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Peking University filed Critical Peking University
Priority to CN201410086745.4A priority Critical patent/CN103823893A/en
Publication of CN103823893A publication Critical patent/CN103823893A/en
Pending legal-status Critical Current

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
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/335—Filtering based on additional data, e.g. user or group profiles
    • G06F16/337—Profile generation, learning or modification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/31—Indexing; Data structures therefor; Storage structures
    • G06F16/313—Selection or weighting of terms for indexing
    • 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
    • G06F16/951—Indexing; Web crawling techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Computational Linguistics (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Software Systems (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention discloses a user comment-based product search method. According to an information requirement provided by a user, a most related product list is searched by the method through combining product data and is returned to the user; the method comprises the following steps: analyzing the product data to obtain an index database, an affective characteristic database and a comment weight database; performing preprocessing and lexical item expansion on a query string submitted by the user to obtain a query lexical item set; searching products and obtaining final score values thereof; ordering from high to low according to the final score values of the products, and cutting off to obtain the product list. By adopting the method, the search effect can be optimized by the product comment information of the user; meanwhile the validity of the introduced information is ensured by analyzing the reference degree in a comment text; in addition, the production search application range and types queried by the user can be expanded; the method is suitable for the applications of product search, gift recommendation and the like of E-business websites.

Description

A kind of product retrieval method and product retrieval system based on user comment
Technical field
The present invention relates to information retrieval field, relate in particular to a kind of product retrieval method and product retrieval system based on user comment.
Background technology
User retrieves product, is the information requirement providing according to user, retrieves the most relevant product list from product library, returns to user.Prior art mainly adopts based on dividing the search method of face and the search method based on keyword.
Wherein, based on the search method of point face, using each structured message of product as a point face, be divided into several classifications.User's inquiry is carried out to participle, and then product category corresponding to judgement inquiry, retrieves with the form of filtering.Such as, for inquiry " the black mobile phones of 2000 left and right ", system will determine demand object for " mobile phone ", and the scope of " price " is [1500,2500], and " color " is " black ", thereby from product library, filters out corresponding product.Meanwhile, system can further be classified by a point face from qualified product, such as, show " brand " to user, more options such as " sizes ", thus further filter.This method does not consider that other user buys the feedback after product.Such as, user's inquiry is " rakish mobile phone ", the system of point region retrieval is by None-identified.This is that and comprise " beautiful ", the product review information of keyword can not be by Direct Classification like this because the system of point region retrieval depends on the classification of each structured message.Therefore, the method for this point of region retrieval cannot be tackled more actual user's inquiry.
Search method based on keyword is divided glossarial index by the information of product by territory, more different weights is distributed in each territory.Participle is carried out in inquiry for user equally, then utilizes point territory marking of existing retrieval model.The higher product of marking is as a result of preferentially returned to user.For example, existing method is just for the language model of a mixing of different information field structure (the Huizhong Duan that gives a mark, ChengXiang Zhai, Jinxing Cheng, Abhishek Gattani.Supporting Keyword Search in Product Database:A Probabilistic Approach, VLDB2013).This method also has deficiency, and one side user's comment text is also not suitable for the directly object as keyword retrieval, because the quality of comment text is different from confidence level, needs to distinguish and treats; On the other hand, for the key word information of product description, not necessarily derive from certain product itself, some overall equivalence class information can be missed under existing retrieval framework.
Summary of the invention
For solving the problems of the prior art, the invention provides a kind of product retrieval method based on user comment, the information requirement that the method provides according to user, the review information of combination product, from product library, excavate the most relevant product list, return to user, the method can be applicable in reality, is applicable in the application such as product retrieval, gift recommendation of electric business website.
Technical scheme of the present invention is:
A product retrieval method based on user comment, the information requirement that the method provides according to user, by combination product data, retrieves the most relevant product list, returns to user, comprises the steps:
The first step: by product data are carried out to data processing and preparation, be specially: product data are carried out to structured analysis, inverted index establishment, affective characteristics extraction, feature equivalence class structure and comment quality analysis, obtain index data base, affective characteristics database and comment weight database;
Second step: submit queries string, carries out query string pre-service to query string;
The 3rd step: by the equivalence class information in affective characteristics database, above-mentioned pretreated query string is carried out to lexical item expansion, obtain inquiring about lexical item collection;
The 4th step: by comment in weight database scoring carry out comprehensively, utilize the concentrated inquiry lexical item of inquiry lexical item in index data base, to carry out product retrieval, to each product retrieving, by obtaining respectively the score value of product feature data and the score value of product review data, obtain the final score value of each product;
The 5th step: by all products that retrieve, carry out getting and blocking after height sorts according to the final score value of each product, obtain product list, return to user.
In the above-mentioned product retrieval method based on user comment, in the first step, product data comprise product feature data and product review data; Structured analysis is specially the product feature data-switching in the product page of electric business website is become after structural data, and by product feature, with < Property Name, the form of property value > represents; Inverted index establishment is specially the product data that described structured analysis is obtained and divides by product attribute value, and is stored as index data base with a form point territory for inverted index; Affective characteristics extracts and is specially by extracting from product review data, obtains the affective characteristics phrase with Affective Evaluation, and feature equivalence class structure is specially by described affective characteristics phrase and carrys out construction feature equivalence class, forms affective characteristics database; Comment quality analysis is specially by described affective characteristics database and product review data comments on matter quantitative analysis, thus structure comment weight database.
Query string pre-service in second step comprises lexical item cutting and part-of-speech tagging operation.
In the 4th step, the score value of product feature data is to obtain by existing keyword retrieval method; The score value of product review data, is first to obtain score value by existing keyword retrieval method, then described score value and the information in comment weight database is weighted and is obtained; The final score value of each product is by above-mentioned all kinds of scorings are summed up and obtained.Wherein, the weight of the score value of product review data can be searched for daily record by electric business and trained and obtain.
For practical application, the invention provides a kind of product retrieval system based on user comment, this system, take index data base, affective characteristics database and comment weight database as support, comprises user's query manipulation module, inquiry pretreatment module, retrieval module and result output module.Wherein, user's query manipulation module is used for user input query string and submits to; Described inquiry pretreatment module is carried out pre-service for the query string that user is submitted to; Described retrieval module is expanded acquisition inquiry lexical item for query string is carried out to lexical item by the equivalence class information of affective characteristics database, inquiry lexical item again by comment in weight database scoring carry out comprehensively, in index data base, carry out product retrieval, to the product retrieving by calculating the score value of its product feature data and the score value of product review data obtains final score value; Described result output module, for the product retrieving is undertaken getting and blocking after height sorts by its final score value, obtains the result of product list, returns to user.
Beneficial effect of the present invention: the invention provides a kind of product retrieval method based on user comment, the information requirement that the method provides according to user, the review information of combination product is excavated the most relevant product list from product library, returns to user.The method is utilized user's product review information, can optimize retrieval effectiveness; Analyze the degree used for reference in comment text simultaneously, guarantee the validity of introducing information; In addition, also can expand the usable range of product retrieval and the type of user's inquiry.In order to be applied in reality, the invention provides a kind of product retrieval system based on user comment, be applicable to the application such as product retrieval, gift recommendation of electric business website.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet of product retrieval process of the present invention.
Fig. 2 is database, module and the schematic flow sheet that product retrieval system of the present invention comprises.
Embodiment
Below in conjunction with the drawings and specific embodiments, the present invention is described in further detail, but the scope not limiting the present invention in any way.
Product retrieval method provided by the invention is product data based on e-commerce website and carrying out.Product data comprise product feature data and product review data.Product feature data refer to the specially structural data of characteristic information such as mark product type, brand, pattern, parameter; The data such as user comment text, evaluation star of product review data pointer to specific products.The present embodiment carries out data processing and preparation to product data, data processing and preparation mainly comprise that structured analysis, the inverted index to product feature data and product review data creates, affective characteristics extracts, feature equivalence class is constructed and comment quality analysis, obtain index data base, affective characteristics database and comment weight database, by submit queries string, product is retrieved on this basis.
User is carried out to user's inquiry log of product retrieval generation and analyze, can find out, have some users to exist and be similar to " mobile phone of giving old man " such query demand.Utilize existing search method, will be difficult to obtain satisfied result.This is because the applicable crowd of product often may not embody in the characteristic of product.If there is user to mention in comment: " this mobile phone is applicable to old man and uses ", just this product can be associated with " old man " this concept.Similarly, because user's comment term variation is comparatively various, keyword wherein can provide more information.Such as, what keyword " outward appearance ", " color ", " workmanship " were expressed is close concept, but because be not near synonym and cannot being contacted in retrieving, by the analysis to user comment, carry out the structure of topic model, the equivalence class that just likely obtains concept is divided, thereby also can return to out the product that is evaluated as " of excellent workmanship ", " color is pretty good " by user in the time of the such query string of reply " buying a goodlooking mobile phone ".
As mistake! Do not find Reference source., the product retrieval method based on user comment provided by the invention comprises the steps:
The first step: by product data are carried out to data processing and preparation, be specially: product data are carried out to structured analysis, inverted index establishment, affective characteristics extraction, feature equivalence class structure and comment quality analysis, obtain index data base, affective characteristics database and comment weight database;
Second step: submit queries string, comprises the query string pre-service of lexical item cutting and part-of-speech tagging operation to query string;
The 3rd step: by the equivalence class information in affective characteristics database, above-mentioned pretreated query string is carried out to lexical item expansion, obtain inquiring about lexical item collection;
The 4th step: by comment in weight database scoring carry out comprehensively, utilize and inquire about the concentrated inquiry lexical item of lexical item and in index data base, carry out product retrieval, obtain product collection; Each product that the above-mentioned product retrieving is concentrated, obtains respectively the score value of product feature data and the score value of product review data; Wherein, the score value of product feature data, obtains by existing keyword retrieval method; The score value of product review data, is first to obtain score value by existing keyword retrieval method, then this score value and the information in comment weight database is weighted; By above-mentioned all kinds of scorings are summed up, obtain the final score value of each product; The weight of above-mentioned all kinds of scorings can be utilized electric business to search for daily record and be trained.
The 5th step: all products that product is concentrated carry out getting and blocking after height sorts according to the final score value of each product, obtain product list, return to user.
For example, user submit to query string be " the good-looking mobile phone of profile of buying to girlfriend ", this query string is carried out to lexical item cutting and part-of-speech tagging and operates pre-service, after pre-service, be expressed as: { " girlfriend ", " profile: good-looking ", " mobile phone " }.By the equivalence class information in affective characteristics database, being expressed as of the inquiry lexical item collection that above-mentioned query string obtains after equivalence class is expanded: { " girlfriend/girl friend/son's wife/schoolgirl ", " profile: good-looking/profile: attractive in appearance/to do manual work: exquisiteness/outward appearance: beautiful ", " mobile phone " }.In product retrieval process, utilize the concentrated inquiry lexical item of inquiry lexical item in index data base, to carry out product retrieval; Because inquiry lexical item concentrated keyword " mobile phone " is unique type matching word, all types be that the product of " mobile phone " is marked all higher than the scoring of non-type of cell phone product in product feature data; In product review data, the comment that those are mentioned to " send girl friend, very beautiful ", " small and exquisite beautiful attractive in appearance, to be applicable to schoolgirl ", will be endowed more scoring.Last add overall score and also therefore can tend to corresponding mobile phone products; Obtain be thus the close to the users product retrieval list of query demand of product.
In the above-mentioned product retrieval method based on user comment, data processing and preparation in the first step are specially: by product data being carried out to structured analysis, inverted index establishment, affective characteristics extraction, feature equivalence class structure and comment quality analysis, obtain index data base, affective characteristics database and comment weight database, its processing and set-up procedure as shown in Figure 2, are described as follows:
1) structured analysis
The data message of electricity business website generally has the following fact: the one, and all information is shown with HTML form; The 2nd, part-structure information can be blended in the content of same html tag.
For the unification of guarantee information, by structured analysis, the html data-switching of the product page of electric business website is become to analyzable structural data.All product attributes are all with <key, and the form that value> is right represents.Such as, typical product attribute comprises: " trade name ", " type ", " brand ", " price ", " comment list " etc.All text messages all carry out cutting and part-of-speech tagging with participle instrument.Usually, the typical attribute of " comment list " comprising: " reviewer ", " comment time ", " comment star " and " comment content ".
2) inverted index creates
This process is divided the structurize product data after above-mentioned analysis by product attribute value, and is stored as index data base with a form point territory for inverted index.The characteristic information of product is regarded one piece of document P as i, its subdocument R is regarded in the comment list of its correspondence as i={ r i, 1, r i, 2..., r i,k.When index building, first by the subdocument R of each product iregard a group as and carry out index, then by document P iadd in index.
3) affective characteristics extracts
This process will extract the phrase with Affective Evaluation, i.e. affective characteristics from product review data.
The phrase that in comment text, certain segment length is n is made as to p={w 1, w 2..., w n, part of speech information corresponding to each word in this phrase is made as q={t 1, t 2..., t n.In this example, in order to guarantee that the phrase of processing is affective characteristic words, limit t 1=noun, t n=adjective.The phrase that is m by another segment length in comment text is made as p '={ v 1, v 2..., v m, make to meet w 1=v 1and w n=v m.Can add up following several information:
(1) frequency that phrase p occurs at all comment texts, is made as tf p;
(2) frequency that the p ' that phrase p is corresponding occurs in all comment texts, is made as tf p';
The feature scores of phrase p is made as score (p), is calculated and is obtained by following formula:
score ( p ) = tf p - &alpha; &Sigma; p &prime; &Element; S &prime; tf p &prime;
Wherein, α is a parameter controlling coincidence degree, for reducing the score value of shorter words group; S' represents the set of qualified all p'.
For all satisfactory each phrases in comment text, carry out descending sort by the feature scores of its acquisition, more therefrom remove the corresponding p ' of all high score phrase p, extract thus and obtain affective characteristics phrase.
4) feature equivalence class structure
This process utilizes the affective characteristics phrase that above-mentioned extraction obtains to carry out construction feature equivalence class, forms affective characteristics database.
Utilize existing LDA method, the affective characteristics phrase obtaining for above-mentioned extraction can extract by following formula the theme of specific quantity: { z 1, z 2..., z k, and each theme Z ilower arbitrary lexical item w jcondition distribute:
z i：{p(w 1|z i)，p(w 2|z i)，...}
Corresponding lexical item w arbitrarily j, derive from each theme Z idistribution, be made as following formula:
w j：{p(w j|z 1），p(w j｜z 2），...}
A certain affective characteristics phrase is by the generating probability p of each theme kbe following formula:
p k : { &Pi; w j &Element; p p ( w j | z 1 ) , &Pi; w j &Element; p p ( w j | z 2 ) , . . . }
In above-mentioned formula, p (w j| z i) represent given theme z i, from topic model generation lexical item w jprobability.
For each affective characteristics, choose the theme of generating probability maximum as its classification, can complete the equivalence class structure of affective characteristics.The example of equivalence class is as < " battery ", " standby " >, < " price ", " cost performance " > etc.
5) comment quality analysis
This process utilizes affective characteristics database and product review data to comment on matter quantitative analysis, thus structure comment weight database, and during for retrieval, the weighting of each comment provides information.
The user comment information of electricity business website has the following fact:
(1) certain customers do not express anyways in the time of comment.The comment of delivering if any some users only includes expression word, interjection etc.;
(2) viewpoint that certain customers express is comparatively single, does not have a reference.If certain customers are because of personal habits, for any commodity of buying, all use similarly comment: " fine fine ", " helping people to buy, well ";
(3) the extensiveness and intensiveness difference that different comment texts are evaluated.
This process obtains comment weight, the contribution providing in the time retrieving product to distinguish different user review information by commenting on quality analysis.Comment quality analysis is from confidence level and two aspect tolerance of availability.
Confidence level refer to comment the viewpoint that provides and true matching degree.First, for all users that comment is provided, add up the comment that it is delivered, and calculate the similarity between comment between two.Definition sim (r i, r j) be the similar function of two comments, if two comment texts are similar, function returns to 1, otherwise is 0.
Similar function is defined as:
sim ( r i , r j ) = 1 dist ( r i , r j ) > = max ( len ( r i ) , len ( r j ) ) 0 otherwise
Wherein, dist (r i, r j) represent the editing distance function of two comment texts; Len (r i) expression comment r ilength, i.e. lexical item number; Max () represents to get both maximal values.
If in the comment that certain user delivered, there is similar comment in the comment of half at least, and all comments of this user are judged to be insincerely, are all filtered.
Whether availability refers to the object that comment evaluates wider, more specifically.Such as, a same product review has been evaluated product appearance and result of use simultaneously, just more has reference than the comment of a certain feature of independent evaluation.And belonging to equally the equivalence class of certain product feature, " color " this feature is just more concrete than " outward appearance ".Owing to having recorded the frequency of each affective characteristics phrase in affective characteristics database, by inverse document frequency idf p, utilize following formula can calculate certain comment r iavailability:
util ( r i ) = &Sigma; p &Element; r i idf p
Wherein, inverse document frequency idf pthe comment number that represents to occur in comment data collection R this affective characteristics phrase p, obtains by following formula:
idf p = log | R | | { r &Element; R : p &Element; r } |
For practical application, the invention provides a kind of product retrieval system based on user comment, this system, take index data base, affective characteristics database and comment weight database as support, comprises user's query manipulation module, inquiry pretreatment module, retrieval module and result output module; Wherein, user's query manipulation module is used for user input query string and submits to; Inquiry pretreatment module is carried out pre-service for the query string that user is submitted to; Retrieval module is expanded acquisition inquiry lexical item for query string is carried out to lexical item by the equivalence class information of affective characteristics database, inquiry lexical item again by comment in weight database scoring carry out comprehensively, in index data base, carry out product retrieval, to the product retrieving by calculating the score value of its product feature data and the score value of product review data obtains final score value; Result output module, for the product retrieving is undertaken getting and blocking after height sorts by its final score value, obtains the result of product list, returns to user.

Claims (7)

1. the product retrieval method based on user comment, is characterized in that, the information requirement that described method provides according to user, by combination product data, retrieves the most relevant product list, returns to user, comprises the steps:
The first step: by product data are carried out to data processing and preparation, be specially: product data are carried out to structured analysis, inverted index establishment, affective characteristics extraction, feature equivalence class structure and comment quality analysis, obtain index data base, affective characteristics database and comment weight database;
Second step: submit queries string, carries out query string pre-service to query string;
The 3rd step: by the equivalence class information in affective characteristics database, above-mentioned pretreated query string is carried out to lexical item expansion, obtain inquiring about lexical item collection;
The 4th step: by comment in weight database scoring carry out comprehensively, utilize the concentrated inquiry lexical item of inquiry lexical item in index data base, to carry out product retrieval, to each product retrieving, by obtaining respectively the score value of product feature data and the score value of product review data, obtain the final score value of each product;
The 5th step: by all products that retrieve, carry out getting and blocking after height sorts according to the final score value of each product, obtain product list, return to user.
2. the product retrieval method based on user comment as claimed in claim 1, is characterized in that, product data comprise product feature data and product review data described in the first step; Described structured analysis is specially the product feature data-switching in the product page of electric business website is become after structural data, and by product feature, with < Property Name, the form of property value > represents; Described inverted index establishment is specially the product data that described structured analysis is obtained and divides by product attribute value, and is stored as index data base with a form point territory for inverted index; Described affective characteristics extracts and is specially by extracting from product review data, obtains the affective characteristics phrase with Affective Evaluation, and described feature equivalence class structure is specially by described affective characteristics phrase and carrys out construction feature equivalence class, forms affective characteristics database; Described comment quality analysis is specially by described affective characteristics database and product review data and comments on matter quantitative analysis, thus structure comment weight database.
3. the product retrieval method based on user comment as claimed in claim 1, is characterized in that, query string pre-service comprises lexical item cutting and part-of-speech tagging operation described in second step.
4. the product retrieval method based on user comment as claimed in claim 1, is characterized in that, the score value of product feature data is to obtain by existing keyword retrieval method described in the 4th step; The score value of described product review data, is first to obtain score value by existing keyword retrieval method, then described score value and the information in comment weight database is weighted and is obtained; The final score value of described each product is by above-mentioned all kinds of scorings are summed up and obtained.
5. the product retrieval method based on user comment as claimed in claim 4, is characterized in that, the weight of the score value of described product review data is searched for daily record by electric business and trained and obtain.
6. the product retrieval system based on user comment, it is characterized in that, described system, take index data base, affective characteristics database and comment weight database as support, comprises user's query manipulation module, inquiry pretreatment module, retrieval module and result output module.
7. the product retrieval system based on user comment as claimed in claim 6, is characterized in that, described user's query manipulation module is for user input query string and submit to; Described inquiry pretreatment module is carried out pre-service for the query string that user is submitted to; Described retrieval module is expanded acquisition inquiry lexical item for query string is carried out to lexical item by the equivalence class information of affective characteristics database, inquiry lexical item again by comment in weight database scoring carry out comprehensively, in index data base, carry out product retrieval, to the product retrieving by calculating the score value of its product feature data and the score value of product review data obtains final score value; Described result output module, for the product retrieving is undertaken getting and blocking after height sorts by its final score value, obtains the result of product list, returns to user.
CN201410086745.4A 2014-03-11 2014-03-11 User comment-based product search method and system Pending CN103823893A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410086745.4A CN103823893A (en) 2014-03-11 2014-03-11 User comment-based product search method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410086745.4A CN103823893A (en) 2014-03-11 2014-03-11 User comment-based product search method and system

Publications (1)

Publication Number Publication Date
CN103823893A true CN103823893A (en) 2014-05-28

Family

ID=50758957

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410086745.4A Pending CN103823893A (en) 2014-03-11 2014-03-11 User comment-based product search method and system

Country Status (1)

Country Link
CN (1) CN103823893A (en)

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104778371A (en) * 2015-04-21 2015-07-15 天脉聚源(北京)传媒科技有限公司 Method and device for evaluating document content speciality
CN105095396A (en) * 2015-07-03 2015-11-25 北京京东尚科信息技术有限公司 Model establishment method, quality assessment method and device
CN105404627A (en) * 2014-09-11 2016-03-16 阿里巴巴集团控股有限公司 Method and device for determining search result
CN105512333A (en) * 2015-12-28 2016-04-20 上海电机学院 Product comment theme searching method based on emotional tendency
CN105653547A (en) * 2014-11-12 2016-06-08 北大方正集团有限公司 Method and device for extracting keywords of text
CN106056209A (en) * 2016-05-23 2016-10-26 大连理工大学 Lexical item weight query learning method based on recurrent neural network
CN106157043A (en) * 2015-03-24 2016-11-23 联想(北京)有限公司 The processing method of a kind of recommended and electronic equipment
CN106407377A (en) * 2016-09-12 2017-02-15 北京百度网讯科技有限公司 Search method and device based on artificial intelligence
CN106445912A (en) * 2016-08-31 2017-02-22 五八同城信息技术有限公司 Evaluation information processing method and apparatus
CN106528832A (en) * 2016-11-15 2017-03-22 南京明鉴智能科技有限公司 Establishment method for comment indexing database and comment indexing system
CN106547748A (en) * 2015-09-16 2017-03-29 中国移动通信集团公司 The creation method and device of a kind of APP index databases, the method and device of search APP
CN106933864A (en) * 2015-12-30 2017-07-07 中国科学院深圳先进技术研究院 A kind of search engine system and its searching method
CN107180111A (en) * 2017-06-13 2017-09-19 深圳市宇数科技有限公司 A kind of information recommendation method, electronic equipment, storage medium and system
CN107577759A (en) * 2017-09-01 2018-01-12 安徽广播电视大学 User comment auto recommending method
CN109857852A (en) * 2019-01-24 2019-06-07 安徽商贸职业技术学院 A kind of the screening judgment method and system of electric business online comment training set feature
CN110019771A (en) * 2017-07-28 2019-07-16 北京国双科技有限公司 The method and device of text-processing
CN110347910A (en) * 2019-05-28 2019-10-18 成都美美臣科技有限公司 One e-commerce website search result filtering rule
CN110837740A (en) * 2019-10-31 2020-02-25 华中科技大学 Comment aspect opinion level mining method based on dictionary improvement LDA model
CN112364170A (en) * 2021-01-13 2021-02-12 北京智慧星光信息技术有限公司 Data emotion analysis method and device, electronic equipment and medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1996363A (en) * 2006-01-01 2007-07-11 腾讯科技（深圳）有限公司 Information displaying method and system
CN101151607A (en) * 2004-12-14 2008-03-26 谷歌公司 Method, system and graphical user interface for providing reviews for a product
CN101515269A (en) * 2008-02-20 2009-08-26 中国科学院自动化研究所 Method for achieving view search engine ranking
CN101702167A (en) * 2009-11-03 2010-05-05 上海第二工业大学 Method for extracting attribution and comment word with template based on internet
CN101901230A (en) * 2009-05-31 2010-12-01 国际商业机器公司 Information retrieval method, user comment processing method and system thereof
US20110320373A1 (en) * 2010-06-25 2011-12-29 Microsoft Corporation Product conversations among social groups

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101151607A (en) * 2004-12-14 2008-03-26 谷歌公司 Method, system and graphical user interface for providing reviews for a product
CN1996363A (en) * 2006-01-01 2007-07-11 腾讯科技（深圳）有限公司 Information displaying method and system
CN101515269A (en) * 2008-02-20 2009-08-26 中国科学院自动化研究所 Method for achieving view search engine ranking
CN101901230A (en) * 2009-05-31 2010-12-01 国际商业机器公司 Information retrieval method, user comment processing method and system thereof
CN101702167A (en) * 2009-11-03 2010-05-05 上海第二工业大学 Method for extracting attribution and comment word with template based on internet
US20110320373A1 (en) * 2010-06-25 2011-12-29 Microsoft Corporation Product conversations among social groups

Cited By (28)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105404627B (en) * 2014-09-11 2019-07-30 阿里巴巴集团控股有限公司 It is a kind of for determining the method and apparatus of search result
CN105404627A (en) * 2014-09-11 2016-03-16 阿里巴巴集团控股有限公司 Method and device for determining search result
CN105653547A (en) * 2014-11-12 2016-06-08 北大方正集团有限公司 Method and device for extracting keywords of text
CN105653547B (en) * 2014-11-12 2020-08-07 北大方正集团有限公司 Method and device for extracting text keywords
CN106157043A (en) * 2015-03-24 2016-11-23 联想(北京)有限公司 The processing method of a kind of recommended and electronic equipment
CN104778371A (en) * 2015-04-21 2015-07-15 天脉聚源(北京)传媒科技有限公司 Method and device for evaluating document content speciality
CN105095396A (en) * 2015-07-03 2015-11-25 北京京东尚科信息技术有限公司 Model establishment method, quality assessment method and device
CN106547748A (en) * 2015-09-16 2017-03-29 中国移动通信集团公司 The creation method and device of a kind of APP index databases, the method and device of search APP
CN105512333A (en) * 2015-12-28 2016-04-20 上海电机学院 Product comment theme searching method based on emotional tendency
CN106933864A (en) * 2015-12-30 2017-07-07 中国科学院深圳先进技术研究院 A kind of search engine system and its searching method
CN106056209A (en) * 2016-05-23 2016-10-26 大连理工大学 Lexical item weight query learning method based on recurrent neural network
CN106056209B (en) * 2016-05-23 2019-04-19 大连理工大学 A kind of inquiry lexical item weight learning method based on Recognition with Recurrent Neural Network
CN106445912A (en) * 2016-08-31 2017-02-22 五八同城信息技术有限公司 Evaluation information processing method and apparatus
CN106407377A (en) * 2016-09-12 2017-02-15 北京百度网讯科技有限公司 Search method and device based on artificial intelligence
CN106407377B (en) * 2016-09-12 2020-03-03 北京百度网讯科技有限公司 Search method and device based on artificial intelligence
CN106528832A (en) * 2016-11-15 2017-03-22 南京明鉴智能科技有限公司 Establishment method for comment indexing database and comment indexing system
CN107180111A (en) * 2017-06-13 2017-09-19 深圳市宇数科技有限公司 A kind of information recommendation method, electronic equipment, storage medium and system
CN107180111B (en) * 2017-06-13 2019-10-25 深圳市宇数科技有限公司 A kind of information recommendation method, electronic equipment, storage medium and system
CN110019771A (en) * 2017-07-28 2019-07-16 北京国双科技有限公司 The method and device of text-processing
CN110019771B (en) * 2017-07-28 2021-08-13 北京国双科技有限公司 Text processing method and device
CN107577759B (en) * 2017-09-01 2021-07-30 安徽广播电视大学 Automatic recommendation method for user comments
CN107577759A (en) * 2017-09-01 2018-01-12 安徽广播电视大学 User comment auto recommending method
CN109857852A (en) * 2019-01-24 2019-06-07 安徽商贸职业技术学院 A kind of the screening judgment method and system of electric business online comment training set feature
CN109857852B (en) * 2019-01-24 2021-02-23 安徽商贸职业技术学院 Method and system for screening and judging characteristics of E-commerce online comment training set
CN110347910A (en) * 2019-05-28 2019-10-18 成都美美臣科技有限公司 One e-commerce website search result filtering rule
CN110837740A (en) * 2019-10-31 2020-02-25 华中科技大学 Comment aspect opinion level mining method based on dictionary improvement LDA model
CN112364170B (en) * 2021-01-13 2021-06-29 北京智慧星光信息技术有限公司 Data emotion analysis method and device, electronic equipment and medium
CN112364170A (en) * 2021-01-13 2021-02-12 北京智慧星光信息技术有限公司 Data emotion analysis method and device, electronic equipment and medium

Similar Documents

Publication Publication Date Title
CN103823893A (en) User comment-based product search method and system
CN101520784B (en) Information issuing system and information issuing method
CN104899273B (en) A kind of Web Personalization method based on topic and relative entropy
CN105718579B (en) A kind of information-pushing method excavated based on internet log and User Activity identifies
CN104008109B (en) Web information Push Service system based on user interest
CN102253982B (en) Query suggestion method based on query semantics and click-through data
CN101420313B (en) Method and system for clustering customer terminal user group
CN103020293B (en) A kind of construction method and system of the ontology library of mobile application
CN104484431B (en) A kind of multi-source Personalize News webpage recommending method based on domain body
CN106339502A (en) Modeling recommendation method based on user behavior data fragmentation cluster
CN101436186A (en) Method and system for providing related searches
CN103838789A (en) Text similarity computing method
CN103577549A (en) Crowd portrayal system and method based on microblog label
CN103049440A (en) Recommendation processing method and processing system for related articles
CN103324665A (en) Hot spot information extraction method and device based on micro-blog
CN105844424A (en) Product quality problem discovery and risk assessment method based on network comments
Xiao et al. Finding News-topic Oriented Influential Twitter Users Based on Topic Related Hashtag Community Detection.
CN102902806A (en) Method and system for performing inquiry expansion by using search engine
CN104268230B (en) A kind of Chinese micro-blog viewpoint detection method based on heterogeneous figure random walk
CN102419778A (en) Information searching method for discovering and clustering sub-topics of query statement
CN102262670A (en) Cross-media information retrieval system and method based on mobile visual equipment
CN102054029A (en) Figure information disambiguation treatment method based on social network and name context
CN105843796A (en) Microblog emotional tendency analysis method and device
CN103455487A (en) Extracting method and device for search term
CN106156135A (en) The method and device of inquiry data

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20140528