CN106296288A - A kind of commodity method of evaluating performance under assessing network text guiding - Google Patents
A kind of commodity method of evaluating performance under assessing network text guiding Download PDFInfo
- Publication number
- CN106296288A CN106296288A CN201610653518.4A CN201610653518A CN106296288A CN 106296288 A CN106296288 A CN 106296288A CN 201610653518 A CN201610653518 A CN 201610653518A CN 106296288 A CN106296288 A CN 106296288A
- Authority
- CN
- China
- Prior art keywords
- commodity
- text
- evaluation
- evaluation object
- comment
- 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
- 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
- G06Q30/0203—Market surveys; Market polls
-
- 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
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Machine Translation (AREA)
Abstract
The present invention provides the commodity method of evaluating performance under the guiding of a kind of assessing network text.Described method includes: a. is for needing to carry out the commodity of performance evaluation, and whole from Network Capture sold goods evaluate text;B. acquired commodity are evaluated text and carry out pretreatment, including going the work such as stop words, participle, mark part of speech;C. obtain the evaluation object of commodity, the commodity performance characteristic that i.e. client may pay close attention to according to the text of evaluating handled well, and calculate the weight of this evaluation object;D. each the comment to commodity carries out pretreatment, structure evaluation object vector, and is normalized;E. extract the emotion word relevant to evaluation object, and compose feeling polarities, extract negative word, the degree adverb etc. being correlated with emotion word simultaneously and compose weights;F. according to the data obtained, calculate the feeling polarities of every comment, thus obtain the feeling polarities of all evaluation texts.
Description
Technical field
The present invention relates to a kind of commodity method of evaluating performance, the method is the commodity under a kind of assessing network text guides
Can evaluation methodology.
Background technology
Along with developing rapidly of ecommerce, shopping online becomes a kind of universal phenomenon, but owing to shopping online lacks
The weary visual contact to commodity, then consult other people just becomes the important way of consumer's acquisition merchandise news to the comment of commodity
Footpath.
But, current commodity appraisement system is the most not fully up to expectations, and the method that most commodity use is poor in being all
Comment, although also have part commodity to add the parameters such as performance such as material, but the performance characteristic that client is paid close attention to often occurs
Not at the row of evaluation, at this time it is accomplished by the comment going to check other client one by one, excavates oneself required information, and this is pole
Time taking.
For trade company, it is also desirable to can the information contained behind of degree of depth digging user evaluation, such as client more pays close attention to
Which performance characteristic of commodity, and existing product do what preferable, having again which aspect is that client payes attention to, and
Shortcomings.
Commodity method of evaluating performance under a kind of assessing network text guiding that the present invention proposes, can be by having bought client
Evaluate and automatically obtain the commodity performance characteristic (evaluation object) that client is paid close attention to, it is also possible to calculate the feeling polarities of every comment
With the feeling polarities of all evaluation texts, consumer and trade company can also according to oneself need excavate information needed.
Summary of the invention
In view of existing technical problem, the present invention provides the commodity method of evaluating performance under the guiding of a kind of assessing network text
Solve this problem.
Including:
A. for needing to carry out the commodity of performance evaluation, whole from Network Capture sold goods evaluate text.
B. acquired commodity are evaluated text and carry out pretreatment, including going the work such as stop words, participle, mark part of speech.
C. obtaining the evaluation object of commodity according to the text of evaluating handled well, the commodity performance that i.e. client may pay close attention to is special
Levy, and calculate the weight of this evaluation object.
D. each the comment to commodity carries out pretreatment, structure evaluation object vector, and is normalized.
E. extract the emotion word relevant to evaluation object, and compose feeling polarities, extract the negative being correlated with emotion word simultaneously
Word, degree adverb etc. also compose weights.
F. according to the data obtained, calculate the feeling polarities of every comment, thus obtain the feeling polarities of all evaluation texts.
Further, in step a, the manufacturer of commodity can obtain the overall network of sold product by agent and comments
Valency text.
Further, assessing network Text Pretreatment to client in step b, including removing stop words and participle, mark word
Property etc. work, this part work Chinese Academy of Sciences Chinese automatic word-cut ICTCLAS or CRF++0.54 can be utilized to carry out.
Further, according to the evaluation text extraction evaluation object handled well in step c, this step is non-the normally off key,
Different from current commodity evaluation methodology, pay attention to the guiding of assessing network text self, i.e. pay close attention to the commodity spy that client is paid close attention to
Property.
Further, needing to compose weight to each evaluation object in step c, this step can be right by calculating each evaluation
The TF-IDF value of elephant, its logarithm value can be as the weight of evaluation object, to represent consumer's attention to this evaluation object
Degree, this method is the commodity evaluation methodology that a kind of commodity evaluate under text guides, and more traditional wants more preferably without bootstrap technique.
Further, in step e, each the comment to commodity carries out extraction emotion word after pretreatment, utilizes sign function
Composing feeling polarities value to emotion word, extract with the relevant negative word of emotion word, degree adverb etc. simultaneously and compose weights, assignment is permissible
Use existing grade scale, it is also possible to combine self-defining method.
Further, according to above the data obtained in step f, calculate the feeling polarities of every comment, comment further according to all
The feeling polarities of opinion calculates the overall emotion polarity of these commodity, consumer and trade company and can also obtain according to the needs of oneself required
Information.
Accompanying drawing explanation
It is right that Fig. 1 is used to illustrate how to extract evaluation in the commodity method of evaluating performance under the guiding of a kind of assessing network text
As, and it is composed the flow chart of weight;
Fig. 2 is used to how illustrate in the inventive method to each comment text structure evaluation object vector, extracts feelings
Sense word is also composed feeling polarities, extraction negative, degree adverb and composes the flow chart of weight.
Detailed description of the invention
Describe the present invention below with reference to detailed description of the invention shown in the drawings.
Fig. 1 is how to extract evaluation object in the commodity method of evaluating performance under assessing network text guides, and composes it
The flow chart of weight.In the present embodiment, mainly comprise the steps that
The all comment text collection selling commodity are collected by agent by trade company.
Collected network comment text set is carried out pretreatment, and the content of pretreatment includes stop words, section in utilization
Institute Chinese automatic word-cut ICTCLAS or CRF++0.54 carries out participle, the work of mark part of speech, and by evaluation literary composition good for pretreatment
Notebook data is derived stand-by;
According to the evaluation text data handled well, extract evaluation object collection O=(o1,o2,…on), and calculate its TF-IDF
Value, composes weight w to evaluation object after taking its natural logrithm1=(w11,w12,…w1n)。
Fig. 2 is how to illustrate in the inventive method to each comment text structure evaluation object vector, extracts emotion word
And compose feeling polarities, extraction negative, degree adverb and compose the flow chart of weight.In the present embodiment, following step is mainly included
Rapid:
For each comment text, carry out pretreatment, include the work such as stop words, participle and mark part of speech equally,
The method taked is consistent with the embodiment in Fig. 1.
To this comment text handled well, construct evaluation object vector V=(v1,v2,…,vn), wherein viRepresent oi?
This comment occurs i time, i=0,1.
In order to prevent feeling polarities from the deviation of many evaluation objects occurring, evaluation object vector need to be normalized P=
(p1,p2,…,pn), wherein
Emotion word Q=(q is extracted from the text that pretreatment is good1,q2,…,qn), and compose feeling polarities value EQ to emotion word
=(e1,e2,…,en), wherein ei=sgn (qiPolarity).
The negative word corresponding with emotion word, degree adverb is extracted from the text that pretreatment is good;If negative word occurs, then
Composing weight-1, the weight of degree adverb refers to existing classification standard tax weight can also compose power in conjunction with self-defining method
Weight after weight, comprehensive negative word and degree adverb is designated as w2=(w21,w22,…w2n)。
According to above the data obtained, calculate the feeling polarities value of this comment:
The feeling polarities of all comment texts is may determine that by the feeling polarities value of all comment texts is averaged
Value.
Additionally, more information can be excavated, such as, by commenting at every of each evaluation object by data above
Proportion in the emotion extreme value of opinion calculates the importance of each evaluation object, it is also possible to extract important evaluation on this basis right
As collection, thus instruct comment on commodity, also provide important references etc. to consumer.
The foregoing is only the better embodiment of the present invention, not in order to limit the present invention, all spirit in the present invention
Within principle, any modification, equivalent substitution and improvement etc. done, should be included within the scope of the present invention.
Claims (7)
1. the commodity method of evaluating performance under assessing network text guides, its feature includes:
A. for needing to carry out the commodity of performance evaluation, whole from Network Capture sold goods evaluate text.
B. acquired commodity are evaluated text and carry out pretreatment, including going the work such as stop words, participle, mark part of speech.C. root
Obtaining the evaluation object of commodity according to the text of evaluating handled well, the commodity performance characteristic that i.e. client may pay close attention to is, and calculate this
The weight of evaluation object.
D. each the comment to commodity carries out pretreatment, structure evaluation object vector, and is normalized.
E. extract the emotion word relevant to evaluation object, and compose feeling polarities, extract negative word, the journey being correlated with emotion word simultaneously
Degree adverbial words etc. also compose weights.
F. according to the data obtained, calculate the feeling polarities of every comment, thus obtain the feeling polarities of all evaluation texts.
2. the commodity method of evaluating performance under assessing network text as claimed in claim 1 guides, is characterized by: by this business
The network agent business of product obtains the whole of sold goods and evaluates text, carries out pretreatment to evaluating text.
3. the commodity method of evaluating performance under assessing network text as claimed in claim 1 guides, is characterized by: evaluate text
Pretreatment be divided into two steps, the first step is stop words, and second step is participle and mark part of speech, and the work of this part can be passed through
Institute of section Chinese automatic word-cut ICTCLAS or CRF++0.54 is carried out.
4. the commodity method of evaluating performance under assessing network text as claimed in claim 1 guides, is characterized by: according to process
Good evaluation text extraction evaluation object, this step is non-the normally off key, different from current commodity evaluation methodology, payes attention to network
Evaluate the guiding of text self, i.e. pay close attention to the product characteristics that client is paid close attention to.
5. the commodity method of evaluating performance under assessing network text as claimed in claim 1 guides, is characterized by: to commodity
Each comment carries out pretreatment, structure evaluation object vector, and is normalized.
6. the commodity method of evaluating performance under assessing network text as claimed in claim 1 guides, is characterized by: to commodity
Each comment carries out extracting after pretreatment emotion word, utilizes sign function to compose feeling polarities value to emotion word, extract simultaneously with
Negative word that emotion word is relevant, degree adverb etc. also compose weights.
7. the commodity method of evaluating performance under assessing network text as claimed in claim 1 guides, is characterized by: according to gained
Data, calculate the feeling polarities of every comment, calculate the overall emotion polarity of these commodity further according to the feeling polarities of all comments,
Consumer and trade company is facilitated to obtain oneself required information.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610653518.4A CN106296288A (en) | 2016-08-10 | 2016-08-10 | A kind of commodity method of evaluating performance under assessing network text guiding |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610653518.4A CN106296288A (en) | 2016-08-10 | 2016-08-10 | A kind of commodity method of evaluating performance under assessing network text guiding |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106296288A true CN106296288A (en) | 2017-01-04 |
Family
ID=57668244
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610653518.4A Pending CN106296288A (en) | 2016-08-10 | 2016-08-10 | A kind of commodity method of evaluating performance under assessing network text guiding |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106296288A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107392718A (en) * | 2017-07-26 | 2017-11-24 | 四川长虹电器股份有限公司 | Method of Commodity Recommendation |
CN108492118A (en) * | 2018-04-03 | 2018-09-04 | 电子科技大学 | The two benches abstracting method of text data is paid a return visit in automobile after-sale service quality evaluation |
CN110096597A (en) * | 2019-03-22 | 2019-08-06 | 同济大学 | A kind of text TF-IDF feature reconstruction method of combination emotional intensity |
CN112506596A (en) * | 2020-11-04 | 2021-03-16 | 北京字节跳动网络技术有限公司 | Object display method and device, electronic equipment and computer readable storage medium |
CN112667969A (en) * | 2020-12-24 | 2021-04-16 | 一汽奔腾轿车有限公司 | Vehicle commodity score calculation method |
WO2021121252A1 (en) * | 2019-12-17 | 2021-06-24 | Beijing Didi Infinity Technology And Development Co., Ltd. | Comment-based behavior prediction |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103365867A (en) * | 2012-03-29 | 2013-10-23 | 腾讯科技(深圳)有限公司 | Method and device for emotion analysis of user evaluation |
CN103885933A (en) * | 2012-12-21 | 2014-06-25 | 富士通株式会社 | Method and equipment for evaluating text sentiment |
-
2016
- 2016-08-10 CN CN201610653518.4A patent/CN106296288A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103365867A (en) * | 2012-03-29 | 2013-10-23 | 腾讯科技(深圳)有限公司 | Method and device for emotion analysis of user evaluation |
CN103885933A (en) * | 2012-12-21 | 2014-06-25 | 富士通株式会社 | Method and equipment for evaluating text sentiment |
Non-Patent Citations (1)
Title |
---|
杨彪: "基于电子商务的评论文本情感极性等级分析", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107392718A (en) * | 2017-07-26 | 2017-11-24 | 四川长虹电器股份有限公司 | Method of Commodity Recommendation |
CN108492118A (en) * | 2018-04-03 | 2018-09-04 | 电子科技大学 | The two benches abstracting method of text data is paid a return visit in automobile after-sale service quality evaluation |
CN108492118B (en) * | 2018-04-03 | 2020-09-29 | 电子科技大学 | Two-stage extraction method for return visit text data of automobile after-sale service quality evaluation |
CN110096597A (en) * | 2019-03-22 | 2019-08-06 | 同济大学 | A kind of text TF-IDF feature reconstruction method of combination emotional intensity |
WO2021121252A1 (en) * | 2019-12-17 | 2021-06-24 | Beijing Didi Infinity Technology And Development Co., Ltd. | Comment-based behavior prediction |
CN112506596A (en) * | 2020-11-04 | 2021-03-16 | 北京字节跳动网络技术有限公司 | Object display method and device, electronic equipment and computer readable storage medium |
CN112667969A (en) * | 2020-12-24 | 2021-04-16 | 一汽奔腾轿车有限公司 | Vehicle commodity score calculation method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106296288A (en) | A kind of commodity method of evaluating performance under assessing network text guiding | |
Zhu et al. | Multimodal c4: An open, billion-scale corpus of images interleaved with text | |
Kim et al. | Sociolinguistic analysis of Twitter in multilingual societies | |
Bühler et al. | Marketing in sports | |
CN102682120B (en) | Method and device for acquiring essential article commented on network | |
Rajeev et al. | Recommending products to customers using opinion mining of online product reviews and features | |
CN104717124A (en) | Friend recommendation method, device and server | |
CN105718526A (en) | Data search method, device and system based on lawyer information | |
Gan et al. | Restaurant Rating: Industrial Standard and Word-of-Mouth--A Text Mining and Multi-dimensional Sentiment Analysis | |
Pujar | Open access journals in library and information science: a study | |
CN102999507A (en) | Recommendation processing method and device for information of network microblog celebrities | |
CN112183078B (en) | Text abstract determining method and device | |
KR20150112089A (en) | Method for servicing recommended goods and apparatus for the same | |
Jafaei Rahni | Explaining and Ranking of the components and indicators of environmental, social and corporate governance reporting by Analytic Hierarchy Process in the companies listed in Stock Exchange | |
CN108170685A (en) | Text emotion analysis method, device and computer readable storage medium | |
Zbierowski | Well-being of entrepreneurs-international comparison based on GEM data | |
KR101994564B1 (en) | Method and apparatus for verifying user | |
CN104731773A (en) | Text sentiment analysis method and text sentiment analysis system | |
KR20150086918A (en) | Device and method for dealing stock using trend sentiment analysis | |
Jamalpur et al. | Opinion mining on restaurant rating based on aspects | |
CN111859146A (en) | Information mining method and device and electronic equipment | |
Yahya et al. | Mobile phone usage among women traders in Tanzania: A case of women traders at Kariakoo Area, Dar-Es-Salaam | |
Heath et al. | Autonomously Communicating Conceptual Knowledge Through Visual Art. | |
Lim et al. | Experience with Mobile Phone Technology: A Comparison Between Two Brands | |
Keerativibool | Forecasting model of monthly gold ornament selling prices |
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 |
Application publication date: 20170104 |
|
WD01 | Invention patent application deemed withdrawn after publication |