CN101833560A - Manufacturer public praise automatic sequencing system based on internet - Google Patents
Manufacturer public praise automatic sequencing system based on internet Download PDFInfo
- Publication number
- CN101833560A CN101833560A CN201010103806A CN201010103806A CN101833560A CN 101833560 A CN101833560 A CN 101833560A CN 201010103806 A CN201010103806 A CN 201010103806A CN 201010103806 A CN201010103806 A CN 201010103806A CN 101833560 A CN101833560 A CN 101833560A
- Authority
- CN
- China
- Prior art keywords
- public praise
- internet
- product
- feature
- evaluation
- 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
Images
Abstract
The invention relates to a manufacturer public praise automatic sequencing system based on internet, overcoming the defect that consumers are difficult to retrieve evaluation information of relative commodities. The invention is used for sequencing public praise for manufacturers and comprises a first sever for receiving requests of on-line visitors, identifying and collecting evaluation information of relative commodities from the internet; a second server for structuralizing and standardizing the collected evaluation information of the relative commodities so as to obtain a public praise sequence for all manufacturers on one commodity; and a third server for distributing the public praise sequencing results for different manufacturers of the relative commodities for the on-line visitors.
Description
Technical field
The present invention relates to a kind of manufacturer public praise automatic sequencing system.
Background technology
The dealing commodity are a kind of regular and important activities, still are that the consumer is significant for the producer.How to make the consumer in the shortest time, recognize that its more comprehensive information of ratio of being concerned about commodity is very important.Be full of in network, product uses a large amount of review information in back to carry out the displaying of structuring and quantification, like product compared all have very important practical value for promoting the producer to understand user's suggestion and consumer.The consumer has bought the product that is fit to regard, generally can compare and understands different manufacturers; And, by also helping finding oneself advantage, inferior position and user's point of interest and preference, thereby provide reference and foundation for production decision with rival's contrast for the producer or businessman.The very big progress that obtains with the information construction of popularizing along with network, increasing people tend to obtain various relevant informations about product by search engine, use back suggestion, producer's prestige and public praise, pricing information etc. as the user, thereby provide support for its consumption decision.In fact, exist a large amount of comment and suggestions on the internet about product.These end users' feedback is more acceptant for the potential buyer with respect to the product introduction webpage that producer provides, and therefore the important references of buying product is provided.
Understand at present problem that the product information approach exists and be can't carry out different product and many features thereof laterally, rapidly, relatively intuitively.General search engine is difficult to carry out special suggestion search, can return a large amount of irrelevant informations etc.Promptly enable accurately to collect the related commentary of relevant a certain product, also can make the user expend more reading time owing to quantity is numerous and jumbled numerous.Existing search engine in search relatively effectively, but in the arrangement of the information relative deficiency that just seems on particularly structuring is handled.Although consumers' opinions is collected by the mode of design of feedback table by a lot of producers, general relative fixed of its list data item and consumer participate in less.The product accounting comparative information confidence level that is designed voluntarily by the producer also is difficult to guarantee in addition.
Product more once had been the work of wasting time and energy very much, and people often understand very few (as parts, function, outward appearance, after sale service, price etc.) to product feature in advance, and in addition, like product often has the selection of multiple different manufacturers, different model.Perhaps the user does not know that certain product all has which brand and manufacturer, does not know to have which related service and relevant auxiliary products.These cause the consumer usually can buy the product of the demand of not meeting, perhaps have use less than function, cause pecuniary waste.Generally speaking, people lack the information of prior overall understanding product feature and operating position.
Summary of the invention
The purpose of this invention is to provide a kind of manufacturer public praise automatic sequencing system, have no way of retrieving defective the dependent merchandise evaluation information to solve the consumer based on the internet.It comprises:
A server is accepted online visitor's request, discerns and collect the evaluation information to dependent merchandise from the internet;
No. two servers carry out structuring and standardization processing to the dependent merchandise evaluation information of collecting, thereby draw the public praise ordering to each manufacturer of same commodity;
Route server is issued the public praise ranking results of the different manufacturers of dependent merchandise to online visitor.
The invention provides the technology of a kind of automatic mining user to the product suggestion.Relying on panoramic product suggestion comment document on the internet is process object, carries out corresponding structureization, quantification treatment, thereby forms public praise comparing result clearly.By comment a large amount of non-structureizations, that expression way is various is changed into clear meaningful regularity structured message, make people come observed data from the angle of macroscopic view.This technology can provide navigation and navigation mechanism, thereby greatly makes things convenient for analysis and decision.Its result can be used for potential user's aid decision making of buying the product purpose for having, also can understand user's reflection and suggestion on the market for generating businessman, and product is done further improvement reference is provided.Characteristics of the present invention are: 1) design quantizes and structured techniques, realizes the fast browsing of public praise contrast; 2) information source makes quantity of information bigger from the internet, and the range of value of information is also very extensive.
Description of drawings
Fig. 1 is a course of work synoptic diagram of the present invention, and Fig. 2 is the evaluation object acquisition methods related with context polarity in embodiment one course of work the 4th step, and Fig. 3 is the synoptic diagram that quantizes bivariate table in the embodiment two.
Embodiment
Embodiment one: specify present embodiment below in conjunction with Fig. 1.Present embodiment comprises:
A server is accepted online visitor's request, discerns and collect the evaluation information to dependent merchandise from the internet;
No. two servers carry out structuring and standardization processing to the dependent merchandise evaluation information of collecting, thereby draw the public praise ordering to each manufacturer of same commodity;
In No. two servers, construct a general emotion tendency dictionary of certain scale; Carry out the product of digging user suggestion at needs, construct corresponding professional polarity speech dictionary; Obtain and discern the sentence of some these products of evaluation; The comment sentence is carried out syntactic analysis, on this basis the corresponding relation of the object of identification and evaluation, evaluation object and viewpoint; According to the recognition result of all sentences, utilize the parts and the attributive character of this product, obtain the result directly perceived of structuring and quantification.
Route server is issued public praise (a certain quality index of promptly popular product to this manufacturer production or the overall price/performance ratio) ranking results of the different manufacturers of dependent merchandise to online visitor.
Work engineering of the present invention is as follows:
One, the accurate identification of product review information with in time obtain comprehensively.Utilize that the product review field is determined relatively, the relative definite characteristics with feature of theme, the relevant language material of large-tonnage product comment is collected in the branch field in advance, it is right therefrom to count the feature word set or the pattern that can be used to describe product review, and periodic retrieval network or pay close attention to some websites, be used for initiatively finding relevant information.Utilize the network themes crawler technology and artificially collect arrangement to combine, obtain the product review corpus of multi-field, the different themes of enough scales; The product review corpus that obtains is carried out topic ratings by importance, and for each theme manually mark the product review keyword and keyword patterns right.Artificial mark keyword and the right effect of keyword patterns can be utilized universal search engine or design independent evaluation system and estimate.If a certain keyword and keyword patterns are then thought to mark successfully to being transfused to universal search engine or after designing independent evaluation system, can returning desired information.And the keyword and the right scoring of keyword patterns of artificial mark also can quantize by its quality and quantity that returns desired information.Utilize the keyword and the keyword patterns of product review corpus and artificial mark right, the keyword of artificial mark and keyword patterns to carrying out the importance evaluation, and are carried out processing such as redundancy elimination, thereby obtain final product review feature set.Utilize final product review feature set retrieval network, obtain potential product review information document flow.
Two, product review information Fast Classification and cluster arrangement.
The tissue of product review information is very crucial with arrangement.Text classification and text cluster are two kinds of very important technology, and the two replenishes mutually.Text classification is a kind of according to experience taxonomic hierarchies and the means of the historical result who trains to information organization, and its classification system is elder generation's group, systematic, has relative independentability between classification and the document.Thereby text classification be fit to according to prediction scheme and domain experts' experience knowledge to product review information type, character and under roughly the judging rapidly of field, help to reduce the search volume.
Than text classification, text cluster then is after text message is arranged earlier class to be arranged, and the character of class and whole classification system are determined by the product review information content that needs are handled fully; On the classification forming process, classification be from always to branch, cluster is total from assigning to.By contrast, clustering processing is refinement more, more can find every aspect, new thread, new theme and even the various rumour of product review information, usual easy uncared-for message subject will be identified after cluster, thereby easier being found, because cluster provides a kind of product review information directly perceived, visual to check and management method.Clustering system need reach higher efficient under the prerequisite that reaches certain cluster quality.Consider the actual features of product review information in addition, cluster need have the ability of handling dynamic text, because obtaining of information is progressive, quantity of information will increase gradually along with going forward one by one of time.As seen, in order to adapt to the requirement of product review to information processing, clustering system should have stronger dynamic self-adapting ability, fast throughput and be convenient to retrieval.
Three, consumer's focus feature identification of product review information.
So-called focal characteristics, emphasis is meant by feature consumer's extensive concern, that can reflect and describe the product review every aspect.Product feature is divided into 5 kinds, i.e. the feature of attribute, parts, parts, related notion, associated components etc.The meaning of extracting focal characteristics also is and can calculates attitude and the suggestion of consumer to most of features by the means of language analysis.Thereby can realize the parametrization of feature, and further calculate the value of its quantification.Therefore the quantification of focal characteristics and corresponding public's suggestion is the important indicator and the description of reflection product review.
The information of delivering from network such as comment suggestion relevant with product review is extracted focal characteristics and can be considered to utilize heuristic rule.Can comprise for the information of utilizing this moment: 1) focal characteristics generally is noun or noun phrase; 2) the general and viewpoint speech of focal characteristics is in certain lexical scope, and the sentence unit of processing can be sentence, simple sentence or chunk; 3) focal characteristics generally has some specific syntactic structures.As " adj (adjective), n (noun) ", " n adj " " isotype; Above-mentioned heuristic rule need be used in combination, for example feature is generally noun or noun phrase, but can not be simply will appear at noun in the user comment as feature, syntax rule information and statistical information that this just need utilize focal characteristics to occur can obtain the syntactic structure of sentence by syntactic analysis for this reason.Can be with noun as candidate feature and carry out probability estimate and estimate its possibility as focal characteristics.If the frequency of a noun in the comprehensive corpus (was example with People's Daily in 1998) of balance is higher in addition, then can reduce its probability as feature.Except above-mentioned rule-based method, can consider to utilize same existing statistical information to find feature with the viewpoint speech.In user comment suggestion language material, viewpoint and attitude are often at a certain concrete feature (variable), for this reason, can utilize the syntactic analysis means, on the artificial mark language material of certain scale, find the same existing pattern and the scale of feature and viewpoint speech, and then utilize these patterns and the viewpoint speech that has the attitude tendency to find feature.
Focal characteristics is to extract according to concrete product, is based on content, and dissimilar products its characteristic parameters is also different, and reaches new technology, product improvement etc. as time passes and change and also will change at its characteristic parameter collection of different phase.Therefore feature extraction mechanism has self-adaptation, dynamic, characteristics such as content-based.In addition, characteristic evaluating relates to the extraction evaluation problem of sentence level and the evaluation problem that general characteristic extracts.Although the feature extraction accuracy of identification of simple sentence is important element task for obtaining of feature set, the obtaining of global feature collection in fact can be tolerated the error of simple sentence feature extraction identification in to a certain degree.
Four, the right automatic extraction of the concern feature of product review information and viewpoint.
The way that adopts the polarity speech to drive is handled: (1) utilizes polarity speech dictionary, at first obtains the polarity speech in the sentence; (2) part of speech of judgement polarity speech, polarity speech part of speech roughly has following several situation: adjective a (for example: " * * * * * is true/d is pretty good/a "), verb v are (for example, " * * * * * comparison/d hommization/v "), noun n (for example, " * * * * * very/d have/v characteristic/n ") etc.; (3) if the polarity speech directly as the modifier of feature, for example " counter-measure timely and effectively ", it is right then can directly to extract feature-viewpoint; (4) if polarity speech part of speech is " a (adjective) ", then need to search forward, search the subject feature of its description, for example can be defined as feature-viewpoint in view of the above right for " camera function " and " hommization " in " camera function seem comparison hommization "; (5) if the polarity speech is " v (verb) ", have two kinds of situations: a kind of is the pattern of band object, verb directly can be composed the core noun of giving in the object; Another situation is the situation that does not have object, and can directly pass to subject with the verb polarity on the predicate this moment; (6) if the polarity speech is " n (noun) ", it is right then directly to extract feature-viewpoint.
The extraction that feature-viewpoint is right be unable to do without the support of tendency dictionary.Can use for reference at present with the relevant Chinese resource of utilizing and comprise the domestic net first published etc. of knowing.But some speech can't strictness provide its polarity, because its attitude direction is dynamic, promptly depends on context environmental and feature, and under different contextual models, its polar orientation is often different.For example, speech such as " greatly ", " little ", " height ", " low ".Though these speech scales are limited, its frequency of occurrences is high especially, almost can be used to modify all features.In order to discern it under different contextual models, the polar orientation during with concrete feature collocation, adopt following method: (1) manually finds these dynamic polarity speech; (2) in the language material scale enough and determine, with the concrete corresponding feature-viewpoint of the feature prerequisite bigger to quantity size under, the public praise of this feature should be stable, can identify its general comment tendency, is made as ψ.If this feature public dispute is bigger, then choose language material again.(3) find the dynamic polarity speech maximum with this feature description; (4) direction of ψ is directly composed the dynamic polarity speech of giving this feature of modification.
Five, the quantification of product review information.
Adopt constant duration that language material is sampled, define the timing node of unequal interval, thereby assess the influence of the generation of subevent (as product improvement and update) subsequent sequence according to the time of origin of product correlator incident.Order is for certain feature dimensions F
yAt sampling time node T
xAnd the attitude information quantization value in the time period between the previous timing node is Attitude (T
x, F
y), this moment, hypothesis had n feature and viewpoint to being identified, and n is wherein arranged
+Individual forward, n
-Individual oppositely (n
++ n
-≤ n), then can calculate it and quantize value, for example by formula Attitude (T
x, F
y)=(n
+-n
-)/n calculates, and can further revise in practice, and for example the n of requirement comment must be greater than certain threshold value.Because for certain feature, if quantized value is approaching zero, then expression is unmanned estimates, and also possible pricer is more, but it is bigger to dispute on.In addition, at identical quantification value (Attitude (T for example
x, F
y)=ξ ') under the situation, then may be the comment of having only the individual same tendency of ξ ', also might be a lot of comments but the difference of the two is ξ '.Can set a confidence factor ψ (0＜ψ≤1) this moment, and its concrete value obtains by the number of times of estimating is carried out staging treating, and it is many more to estimate number of times, and the degree of confidence that quantizes value is also big more.Then Ci Shi formula will be modified to Attitude (T
x, F
y)=ψ (n
+-n
-)/n.Make that bigger its quantification value of feature of degree of confidence is strengthened.
The effect that quantizes is: 1) by big quantitative statistics, can the feature that cause extensive concern is outstanding, and feature is carried out Classification Management or carried out the feature importance ranking; 2) evaded because the problem that the Language Processing technology acuracy is not high, individual statements does not meet syntax gauge and possible error is brought.Quantize to be based upon on the right identification basis of foregoing feature and viewpoint, the data after the quantification can form more accurateization and correlation data intuitively on different feature dimensions and timing node.
Embodiment two: the difference of present embodiment and embodiment one is: No. one server produces the comment language material towards specific products that is used for getting access to enough scales behind the query expression of internet retrieval product review information earlier;
No. two servers to the comment language material of specific products carry out that evaluation object obtains, feature-viewpoint is to processing such as identifications, particularly the polarity problems of value when related with different characteristic for dynamic polarity speech has been taked a kind of effective processing policy; On this basis, carry out structuring and quantize computing, thereby obtain the different vendor of this series products and the quantification bivariate table of model, and the compare operation of sorting in view of the above, thereby for consumers and producers's reference.
The quantification bivariate table is a bivariate table, and wherein one dimension is the different vendor and the model of product, and one dimension is all features of this kind product in addition.For example, for mobile phone products, feature comprises " battery ", " feel ", " camera function " etc.Data in the bivariate table be the user to same feature, the public praise quantized values of different model product.
Claims (3)
1. based on the manufacturer public praise automatic sequencing system of internet, it is characterized in that it comprises:
A server is accepted online visitor's request, discerns and collect the evaluation information to dependent merchandise from the internet;
No. two servers carry out structuring and standardization processing to the dependent merchandise evaluation information of collecting, thereby draw the public praise ordering to each manufacturer of same commodity;
Route server is issued the public praise ranking results of the different manufacturers of dependent merchandise to online visitor.
2. the manufacturer public praise automatic sequencing system based on the internet according to claim 1 is characterized in that constructing a general emotion tendency dictionary of certain scale in No. two servers; Carry out the product of digging user suggestion at needs, construct corresponding professional polarity speech dictionary; Obtain and discern the sentence of some these products of evaluation; The comment sentence is carried out syntactic analysis, on this basis the corresponding relation of the object of identification and evaluation, evaluation object and viewpoint; According to the recognition result of all sentences, utilize the parts and the attributive character of this product, obtain the result directly perceived of structuring and quantification.
3. the manufacturer public praise automatic sequencing system based on the internet according to claim 1 is characterized in that a server produces the comment language material towards specific products that is used for getting access to enough scales behind the query expression of internet retrieval product review information earlier; No. two servers to the comment language material of specific products carry out that evaluation object obtains, feature-viewpoint is to processing such as identifications; Carry out structuring on this basis and quantize computing, thereby obtain the different vendor of this series products and the quantification bivariate table of model, and the compare operation of sorting in view of the above.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010103806A CN101833560A (en) | 2010-02-02 | 2010-02-02 | Manufacturer public praise automatic sequencing system based on internet |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010103806A CN101833560A (en) | 2010-02-02 | 2010-02-02 | Manufacturer public praise automatic sequencing system based on internet |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101833560A true CN101833560A (en) | 2010-09-15 |
Family
ID=42717631
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010103806A Pending CN101833560A (en) | 2010-02-02 | 2010-02-02 | Manufacturer public praise automatic sequencing system based on internet |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101833560A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102841946A (en) * | 2012-08-24 | 2012-12-26 | 北京国政通科技有限公司 | Commodity data retrieval sequencing and commodity recommendation method and system |
| CN103235788A (en) * | 2013-03-28 | 2013-08-07 | 北京百度网讯科技有限公司 | Method and equipment used for obtaining evaluation information result on target object |
| CN103577542A (en) * | 2013-10-10 | 2014-02-12 | 北京智谷睿拓技术服务有限公司 | Ranking fraud detection method and ranking fraud detection system of application program |
| CN103870973A (en) * | 2012-12-13 | 2014-06-18 | 阿里巴巴集团控股有限公司 | Information push and search method and apparatus based on electronic information keyword extraction |
| CN104484815A (en) * | 2014-12-18 | 2015-04-01 | 刘耀强 | Product-oriented emotion analysis method and system based on fuzzy body |
| CN104657425A (en) * | 2014-10-06 | 2015-05-27 | 中华电信股份有限公司 | Topic management type network public opinion evaluation management system and method |
| CN104731903A (en) * | 2015-03-23 | 2015-06-24 | 魏强 | Method for searching for enterprise on basis of products and search device |
| CN106021433A (en) * | 2016-05-16 | 2016-10-12 | 北京百分点信息科技有限公司 | Public praise analysis method and apparatus for product review data |
| CN106294355A (en) * | 2015-05-14 | 2017-01-04 | 阿里巴巴集团控股有限公司 | A kind of determination method and apparatus of business object attribute |
| CN107369066A (en) * | 2017-06-28 | 2017-11-21 | 东软集团股份有限公司 | A kind of feature between comment object compares method and device |
| CN107729564A (en) * | 2017-11-13 | 2018-02-23 | 北京众荟信息技术股份有限公司 | A kind of distributed focused web crawler web page crawl method and system |
| CN108885750A (en) * | 2016-04-05 | 2018-11-23 | 赫尔实验室有限公司 | Emerging defect and safety monitoring system |
| CN110991838A (en) * | 2019-11-21 | 2020-04-10 | 中国联合网络通信集团有限公司 | Method and device for determining competitiveness index of communication operator |
-
2010
- 2010-02-02 CN CN201010103806A patent/CN101833560A/en active Pending
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102841946A (en) * | 2012-08-24 | 2012-12-26 | 北京国政通科技有限公司 | Commodity data retrieval sequencing and commodity recommendation method and system |
| CN103870973A (en) * | 2012-12-13 | 2014-06-18 | 阿里巴巴集团控股有限公司 | Information push and search method and apparatus based on electronic information keyword extraction |
| CN103870973B (en) * | 2012-12-13 | 2017-12-19 | 阿里巴巴集团控股有限公司 | Information push, searching method and the device of keyword extraction based on electronic information |
| CN103235788A (en) * | 2013-03-28 | 2013-08-07 | 北京百度网讯科技有限公司 | Method and equipment used for obtaining evaluation information result on target object |
| CN103577542A (en) * | 2013-10-10 | 2014-02-12 | 北京智谷睿拓技术服务有限公司 | Ranking fraud detection method and ranking fraud detection system of application program |
| CN104657425A (en) * | 2014-10-06 | 2015-05-27 | 中华电信股份有限公司 | Topic management type network public opinion evaluation management system and method |
| CN104657425B (en) * | 2014-10-06 | 2019-02-22 | 中华电信股份有限公司 | Topic management type network public opinion evaluation management system and method |
| CN104484815B (en) * | 2014-12-18 | 2017-11-21 | 刘耀强 | Based on fuzzy ontology towards the sentiment analysis method and system in terms of product |
| CN104484815A (en) * | 2014-12-18 | 2015-04-01 | 刘耀强 | Product-oriented emotion analysis method and system based on fuzzy body |
| CN104731903A (en) * | 2015-03-23 | 2015-06-24 | 魏强 | Method for searching for enterprise on basis of products and search device |
| CN106294355A (en) * | 2015-05-14 | 2017-01-04 | 阿里巴巴集团控股有限公司 | A kind of determination method and apparatus of business object attribute |
| CN108885750A (en) * | 2016-04-05 | 2018-11-23 | 赫尔实验室有限公司 | Emerging defect and safety monitoring system |
| CN106021433A (en) * | 2016-05-16 | 2016-10-12 | 北京百分点信息科技有限公司 | Public praise analysis method and apparatus for product review data |
| CN106021433B (en) * | 2016-05-16 | 2019-05-10 | 北京百分点信息科技有限公司 | A kind of the public praise analysis method and device of comment on commodity data |
| CN107369066A (en) * | 2017-06-28 | 2017-11-21 | 东软集团股份有限公司 | A kind of feature between comment object compares method and device |
| CN107369066B (en) * | 2017-06-28 | 2021-05-28 | 东软集团股份有限公司 | Feature comparison method and device between comment objects |
| CN107729564A (en) * | 2017-11-13 | 2018-02-23 | 北京众荟信息技术股份有限公司 | A kind of distributed focused web crawler web page crawl method and system |
| CN110991838A (en) * | 2019-11-21 | 2020-04-10 | 中国联合网络通信集团有限公司 | Method and device for determining competitiveness index of communication operator |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101833560A (en) | Manufacturer public praise automatic sequencing system based on internet | |
| CN109815308B (en) | Method and device for determining intention recognition model and method and device for searching intention recognition | |
| Hu et al. | Opinion mining from online hotel reviews–a text summarization approach | |
| Keneshloo et al. | Predicting the popularity of news articles | |
| TWI557664B (en) | Product information publishing method and device | |
| TWI615724B (en) | Information push, search method and device based on electronic information-based keyword extraction | |
| CN106649603B (en) | Designated information pushing method based on emotion classification of webpage text data | |
| US20180052823A1 (en) | Hybrid Classifier for Assigning Natural Language Processing (NLP) Inputs to Domains in Real-Time | |
| Sutabri et al. | Improving naïve bayes in sentiment analysis for hotel industry in Indonesia | |
| CN104199833B (en) | The clustering method and clustering apparatus of a kind of network search words | |
| CN108038725A (en) | A kind of electric business Customer Satisfaction for Product analysis method based on machine learning | |
| CN109766431A (en) | A kind of social networks short text recommended method based on meaning of a word topic model | |
| CN104978665A (en) | Brand evaluation method and brand evaluation device | |
| CN103324665A (en) | Hot spot information extraction method and device based on micro-blog | |
| CN110263248A (en) | A kind of information-pushing method, device, storage medium and server | |
| Wang et al. | Customer-driven product design selection using web based user-generated content | |
| CN105426514A (en) | Personalized mobile APP recommendation method | |
| CN107544988B (en) | Method and device for acquiring public opinion data | |
| CN104268292A (en) | Label word library update method of portrait system | |
| CN103559207A (en) | Financial behavior analyzing system based on social media calculation | |
| KR20120108095A (en) | System for analyzing social data collected by communication network | |
| CN105069103A (en) | Method and system for APP search engine to utilize client comment | |
| Ojokoh et al. | A feature-opinion extraction approach to opinion mining | |
| CN105468649A (en) | Method and apparatus for determining matching of to-be-displayed object | |
| Guo et al. | A survey of internet public opinion mining |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20100915 |