CN110706028A - Commodity evaluation emotion analysis system based on attribute characteristics - Google Patents
Commodity evaluation emotion analysis system based on attribute characteristics Download PDFInfo
- Publication number
- CN110706028A CN110706028A CN201910916447.6A CN201910916447A CN110706028A CN 110706028 A CN110706028 A CN 110706028A CN 201910916447 A CN201910916447 A CN 201910916447A CN 110706028 A CN110706028 A CN 110706028A
- Authority
- CN
- China
- Prior art keywords
- attribute
- commodity
- emotion
- evaluation
- word
- 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
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
- G06Q30/0282—Rating or review of business operators or products
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- Databases & Information Systems (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- General Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of big data, and discloses a commodity evaluation emotion analysis system based on attribute characteristics, which is used for acquiring emotional tendency of a consumer to specific commodities and different attributes of the commodities so as to provide reference for research, development and sale of subsequent products by purchasing and evaluating. The invention comprises the following steps: the user interaction module is used for authenticating the identity of the user and receiving the website of the target commodity input by the user; the evaluation statement acquisition module is used for capturing evaluation data of the consumer; the data preprocessing module is used for preprocessing data; the extraction module is used for extracting the attribute characteristics and the attribute weight of the commodity from the preprocessed data to obtain attribute characteristic-emotion word pairs; the algorithm processing module is used for inputting the extracted attribute feature-emotion word pair into an algorithm model for computational analysis, realizing an emotion analysis algorithm based on the attribute feature and outputting a commodity evaluation emotion quantitative value; and the result display module is used for displaying the analysis result. The invention is suitable for commodity evaluation emotion analysis.
Description
Technical Field
The invention relates to the technical field of big data, in particular to a commodity evaluation emotion analysis system based on attribute characteristics.
Background
The commodity evaluation is subjective feeling of using the commodity by the consumer, can express the emotional tendency of the consumer, and has wider experience range and more experience number compared with the commodity information of the merchant. Due to the rise of electronic commerce, consumers can buy self-mind commodities without leaving home, the popularization of Web2.0 makes it a habit that consumers publish commodity evaluation on the internet after buying commodities, various unstructured text contents in the internet are increased explosively, and on the basis of the habit, research on evaluation text emotion tendency analysis based on Natural Language Processing (NLP) technology is carried forward.
In the face of increasingly complex text data and increasing text emotion analysis requirements, the existing text emotion analysis system has the following defects:
(1) the acquisition, processing and analysis processes are mutually independent, and the integration of the whole stage is not realized;
(2) the data set source platform is single, the workload of acquiring data is large, and the repeatability is high;
(3) the emotion analysis result often lacks a uniform and quantitative standard, and a user needs to further process the analysis result to obtain a data result and a statistical chart which can be compared;
(4) in the existing online evaluation emotion research, students often regard online comments as atomic objects, and the research on evaluation emotion stays at the level of a single product, but lacks consideration of evaluation emotion of different attributes of the same product with finer granularity. For a small amount of emotional researches based on commodity attributes, attributes and attribute weights are often set artificially, and objective fairness of evaluation results is difficult to guarantee.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the commodity evaluation emotion analysis system based on the attribute features is used for obtaining emotional tendency of a consumer to specific commodities and different attributes of the commodities so as to provide references for research, development and sale of subsequent products for purchasing and evaluation.
In order to solve the problems, the invention adopts the technical scheme that: the commodity evaluation emotion analysis system based on the attribute characteristics comprises the following modules:
a user interaction module: used for authenticating the identity of the user, receiving the website of the target commodity input by the user and transmitting the website to the system background,
an evaluation statement acquisition module: capturing commodity details and consumer evaluation data from a website of a target commodity by using a crawler technology;
a data preprocessing module: the evaluation statement acquisition module is used for acquiring evaluation statements of the user;
an extraction module: the system is used for extracting the attribute characteristics and the attribute weight of the commodity from the preprocessed data, and obtaining emotional words matched with the attribute characteristics from the consumer evaluation data through semantic relation calculation, word segmentation and part-of-speech tagging methods so as to obtain attribute characteristic-emotional word pairs;
an algorithm processing module: the system is used for inputting the extracted attribute feature-emotion word pair into an algorithm model for computational analysis, realizing an emotion analysis algorithm based on the attribute feature and outputting a commodity evaluation emotion quantitative value;
and a result display module: and displaying the analysis result.
Further, the preprocessing operation of the data preprocessing module may include: word segmentation, filtering, part of speech tagging and word frequency statistics.
Further, the extracting module extracts the attribute features of the commodity, including the explicit attribute feature and the implicit attribute feature, and the specific steps of extracting the attribute features include:
determining the explicit attribute characteristics of the commodity according to the commodity details and the consumer evaluation information;
screening out all nouns and nominal phrases in the comment set as a candidate word set through the statistical word frequency;
through point mutual information calculation, nouns or nominal phrases with high mutual information values with the explicit attribute characteristics in the candidate word set are identified and used as the implicit attribute characteristics, and finally the attribute characteristic set of the commodity is formed by the explicit attribute characteristics and the implicit attribute characteristics.
Further, the formula of the point-to-point information calculation may be as follows:
wherein, PMI (F)iPh) as an explicit Attribute feature FiName of HeMutual information value of word or noun phrase ph, ph being noun or noun phrase in comment set, p (F)iPh) explicit Attribute feature F for comment setsiProbability of co-occurrence with noun or noun phrase ph, p (F)i) Aggregating explicit attribute features for comments FiThe probability of occurrence, p (ph), is the probability of occurrence of the candidate attribute word in the comment set.
Furthermore, the extraction module can obtain the attribute weight of the commodity according to the word frequency statistical result.
Further, the mode of the algorithm processing module for implementing the emotion analysis algorithm based on the attribute features may be as follows:
single set of attribute feature-emotion word pairs<Feature(i),Opinion(j)>Sentiment polarity base score of Sen _ scoreijIs defined as:
then the emotion scoring algorithm for the individual attribute features of a certain commodity is as follows:
the sentiment score of a single good is formulated as follows:
wherein m represents the number of attribute features contained in a single commodity, n is the number of emotion words appearing in the attribute features,for the degree of adverb intensity level, phi, occurring before each emotional wordDegreeCoefficient values corresponding to the strength levels of the adverbs with corresponding degrees, f is the number of negative words before the emotional word, mu is the score value of the negative word of the emotional word, class1-class4 are the strength levels of the adverbs with four divided degrees, and alphaiIs the weight coefficient of the attribute feature, i is the serial number of the attribute feature, i belongs to {1,2,3, … …, m }, betaijSentiment polarity base score, beta, for the jth sentiment word of the ith attribute featureijThe value range is [ -1,1 [ ]]Wherein positive and negative correspond to positive and negative emotions, respectively.
Further, the analysis results displayed by the result display module may include: the method comprises the following steps of commodity attribute scoring, radar chart scoring of each attribute of the commodity, total commodity scoring based on attribute weight and system running conditions.
The invention has the beneficial effects that: the invention can effectively collect and process commodity evaluation information from different platforms, can help merchants to quickly and widely collect feedback of consumers to products, including use experience of competitive product consumers, and is beneficial to the merchants to adjust research and development directions according to the use experience of the consumers.
Drawings
FIG. 1 is a flow chart of a commodity evaluation sentiment analysis system based on attribute characteristics according to the present invention;
FIG. 2 is a flow diagram of a background module of the system provided by the present invention;
FIG. 3 is a combined emotion dictionary structure diagram constructed by the present invention.
Detailed Description
In order to obtain the emotional tendency of a consumer to specific commodities and different attributes of the commodities by processing a large amount of unstructured commodity evaluation data on the premise of ensuring efficiency and accuracy, so that the purchase evaluation provides reference for research, development and sale of subsequent products to the maximum extent.
The invention provides a commodity evaluation emotion analysis system based on attribute characteristics, which is structurally shown in figure 1 and comprises the following system modules:
the system comprises a first module, namely a user interaction module, a user interaction module and a back-end module, wherein the first module is a user interaction module, the user interaction module is used for ① to authenticate the user identity in the login process and assign different authorities to users with different identities, so that the users with different identities enter respective interfaces to obtain the interface presentation required by the users, and ② receives the commodity website input by the users and transmits the commodity website to a system back-end.
And a second module: namely an evaluation statement acquisition module. The method is used for automatically acquiring commodity evaluation data sets required by emotion analysis. According to the invention, a crawler technology is embedded in the system, the commodity evaluation data set can be automatically crawled after the commodity website input by the user is received, and the user does not need to obtain the evaluation data from other ways and then upload the evaluation data.
And a third module: namely a data pre-processing module. The method comprises the steps of evaluating whether stop words which influence the text analysis efficiency are existed in the text, wherein the stop words comprise nonsense auxiliary words such as 'and' or 'and', and functional symbols such as '@', '/', '#', and the like which are irrelevant to the text content analysis. In addition, the commodity evaluation data set also comprises an evaluation which is fixed in structure and does not have the actual meaning of text content analysis, for example, after a user purchases a commodity, the evaluation of the relevant commodity is not made for a long time, and the default evaluation of the Jingdong mart is as follows: "this user does not fill in the evaluation content". Therefore, partial stop words and meaningless evaluation texts need to be filtered, so that the storage space is saved, the text processing efficiency is improved, and the accuracy of the processing result is ensured. And the accurate extraction of the attribute words and the emotional conditions in the evaluation text is the basis for extracting the attribute feature-emotional word pair in the module IV, so the data preprocessing module also carries out part-of-speech tagging on the word segmentation result. In order to make the display result in the module six more specific and comprehensive, descriptive statistics including data set size, word frequency statistics, evaluation text word cloud pictures and the like also need to be generated in the data preprocessing module.
And a module IV: the module IV can determine the attribute feature set of the commodity according to the crawled product details and the user evaluation information. In part of commodity evaluation texts, consumers indirectly express the satisfaction degree of the evaluators for a certain attribute, the evaluation lacks obvious attribute words and contains rich emotional information of the consumers, so that compared with the existing emotion analysis system which only analyzes and researches the displayed attribute, the extraction of the attribute words in the invention also comprises the extraction of the implicit attribute of the commodity.
The specific steps of the module four for extracting the explicit attribute features and the implicit attribute features can be as follows:
1) and (5) explicit attribute feature extraction.
Because the third module carries out part-of-speech tagging and word frequency statistics on the evaluation information of the consumer, when the explicit attribute characteristics of the commodity are determined, a noun with higher tagged word frequency can be selected to be added into the explicit attribute characteristics, and the attributes which are introduced originally on the detailed page of the commodity are added to form the final explicit attribute characteristics;
2) and extracting the implicit attribute.
Firstly, screening out all nouns and nominal phrases in a comment set as a candidate word set through statistical word frequency, then identifying nouns or nominal phrases with high mutual information value with explicit attribute characteristics in the candidate word set through Point Mutual Information (PMI) calculation, and taking the nouns or nominal phrases as implicit attribute characteristics. The point-to-point information calculation formula is shown in formula (1):
wherein, PMI (F)iPh) as an explicit Attribute feature FiMutual information value with noun or noun phrase ph, ph being noun or noun phrase in comment set, p (F)iPh) explicit Attribute feature F for comment setsiProbability of co-occurrence with noun or noun phrase ph, p (F)i) Aggregating explicit attribute features for comments FiThe probability of occurrence, p (ph), is the probability of occurrence of the candidate attribute word in the comment set. The candidate word is marked once or more times in each comment. The higher the PMI value, the more confident the set of candidate words is put into the attribute set.
3) The explicit attribute features and the implicit attribute features obtained in the steps 1) and 2) can jointly form an attribute feature set of the commodity. In order to improve the attribute feature set, the attribute feature set can be subsequently expanded by combining manual judgment, synonym forest and the like.
In consideration of the difference of importance of the attribute characteristics of the comment objects, the invention introduces an attribute weight influence factor, which is referred to as attribute weight for short. Since the online comments are all short texts, the traditional feature weight method for long texts is not applicable. The design attribute weight calculation method adopts the following basis: if the word frequency of the attribute word is larger and the number of evaluation items containing the attribute word is larger, the importance degree of the attribute word is higher. According to the word frequency statistical result of data preprocessing, the calculated attribute weight coefficient is recorded as alphaiAnd i (i is equal to {1,2,3, … …, m }) is the attribute feature number.
The unstructured evaluation text has no feasibility of analyzing and scoring, so that in the matching process of the attribute features and the emotion words, the mathematical features of products are labeled as F (Feature), the emotion words are labeled as O (Opinion including emotion adjectives and negative words and degree adverbs for modifying the adjectives), the module obtains the emotion words matched with the attribute features from the evaluation data of consumers through semantic relation calculation, word segmentation and part-of-speech labeling methods, and finally obtains the attribute features-emotion word pair < Feature, Opinion >.
And a fifth module: the algorithm processing module is internally provided with an algorithm model and is used for inputting the structured attribute feature-emotion word pair into the algorithm model for computational analysis, realizing the emotion analysis algorithm based on the attribute feature and outputting the emotion quantization value of the comment set. The calculation rule of the text sentiment value is a core element of the text sentiment analysis score, and combines the existing conventional text analysis rule with the characteristics of the commodity evaluation information, so that the calculation rule of the sentiment strong score is constructed as follows:
extraction of<Feature(i),Opinion(j)>The adjectives in the middle Opinion as the emotional words score the emotional polarities contained in the adjectives as betaijThat is, the value range of the sentiment polarity basic score of the jth sentiment word aiming at the ith attribute feature is [ -1,1]Wherein, is positive and negativeCorresponding to positive and negative emotions, respectively, the absolute value is larger the stronger the emotion is.
Meanwhile, the emotion word basic score is multiplied by the corresponding emotion intensity degree weight of the emotion modifying degree adverb; and if negative words exist before the emotional words, counting the occurrence times of all the negative words, and if the negative words are odd, multiplying the negative words by-1 before the emotional polarity basic score of the emotional words. The calculation rule of the emotion value can be finally constructed as:
the attribute feature set of a single commodity contains m attribute features of the commodity, each attribute feature has n emotion words matched with a combined emotion dictionary, the combined emotion dictionary is shown in figure 3, and the degree adverb strength level of each emotion word appearing before is
φDegreeThe coefficient value corresponding to the corresponding degree adverb strength level is shown, f is the number of the negative words before the emotional word, mu is the score of the negative words of the emotional word, class1-class4 are four divided degree adverb strength levels, such as adverbs of 'extraordinary', 'extraordinary' and 'some', and each adverb corresponds to a corresponding adverb strength level.
According to the calculation rule for constructing the emotional words, the invention forms a single group<Feature(i),Opinion(j)>Sentiment polarity base score of Sen _ scoreijDefined by formula (2):
wherein
Adding all emotion word emotion polarity base scores of single attribute characteristics, and dividing the sum by the evaluation numberRelated Property features F in the datasetiTotal number of occurrences of emotional words n (F)i) The emotion value algorithm for obtaining the single attribute feature is as shown in formula (5):
since the final sentiment value of a single commodity needs to consider not only the sentiment value of a single attribute of the commodity but also the difference in importance of different attributes, when calculating the final sentiment value of a single commodity, the sentiment value of a single attribute should be multiplied by the weighting coefficient of the attribute and then summed up, as shown in formula (6):
and a module six: the result display module is used for displaying the currently generated calculation analysis result (such as each attribute score of the commodity, each attribute score radar chart of the commodity or total scores of the commodity based on attribute weights) and the system operation condition to the user. The analysis results of the existing emotion analysis system often lack a uniform quantification standard, and a user is required to further process the analysis results so as to obtain data results and statistical charts which can be compared. The invention integrates the process into the system, and simplifies the operation process of the user.
Combining the functional descriptions of the above modules, the specific workflow of the present invention can be as shown in fig. 2.
Examples
Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings.
The commodity evaluation emotion analysis system based on attribute features provided by the embodiment is combined with the system structure shown in fig. 1 and the work flow shown in fig. 2, and the working principle of the embodiment is as follows:
1. user interaction and ratings data acquisition
After logging in the system, the user inputs a website corresponding to a single target commodity on a platform where the target commodity is located (in this example, a certain mobile phone product in the kyoto mall is taken as an example), and the system obtains the website and transmits the website to the background module. The system background utilizes a crawler technology to crawl consumer evaluation data on different shopping platforms and stores the data into a Mongobb database. The platform comprises a shopping platform and a commenting reservation platform, and specifically comprises Taobao, Techthyst, Jingdong, Mei Tuo, popular commenting and the like.
2. Data pre-processing
The data preprocessing module needs to perform the following steps in sequence: filtering nonsense texts, segmenting words, removing stop words, labeling parts of speech, counting word frequency and evaluating text word cloud pictures to generate.
3. Extracting commodity attribute feature and emotion word pair
And determining an attribute class set according to the crawled product details and the user evaluation information. Taking a mobile phone commodity as an example, the crawling user evaluation information summary result comprises the following attributes of memory, appearance, price, screen and running speed, the commodity details comprise commodity summaries of memory, endurance, screen and the like, and the commodity attribute set can be determined to be an attribute set comprising the following 6 explicit attribute characteristics of memory, appearance, price, endurance, screen and running speed.
The extraction of the attribute words also comprises the extraction of the implicit attribute of the commodity. For example, if a certain mobile phone commodity is evaluated as 'game is not blocked, back shell color is good, but the position of the front camera is too ugly', the first clause of the evaluation is actually the description of the attribute of 'running speed' of the mobile phone product by the consumer, the second clauses both imply the evaluation of the consumer on the appearance of the mobile phone, and the implicit attribute is the 2 nd attribute characteristic 'appearance' in the attribute set and is marked as F2。
And (3) screening out all nouns and nominal phrases in the comment set as a candidate word set through word frequency statistics of a module III, such as { game, backshell, front camera, … … }, and calculating PMI value:
and judging whether the candidate word set is classified into the attribute feature set as an implicit attribute feature according to the PMI value, and finally forming the attribute feature set of the commodity by using the explicit attribute feature and the implicit attribute feature contained in the evaluation data.
The embodiment determines the attribute weight by taking the word frequency of the attribute words as the basis, and records the calculated attribute weight coefficient as alpha according to the word frequency statistical result of data preprocessingi(i∈{1,2,3,……,6})：
And finally, combining semantic relation calculation, word segmentation and part-of-speech tagging, obtaining emotion words matched with attribute features from the evaluation data of the consumers, thereby extracting attribute features-emotion word pairs < feature (i) and opinion (j) in the evaluation data, and storing the attribute features-emotion word pairs in a database.
4. Evaluation text emotion analysis algorithm
Taking the mobile phone product evaluation that the back shell color is good, but the position of the front camera is too ugly as an example, 3 groups of attribute feature-emotion word pairs can be obtained by extracting the implicit attribute, namely<Speed of operation, not jamming>、<The appearance is beautiful>、<Outward appearance, too ugu>. And (4) scoring the emotional words in each group of attribute feature-emotional word pairs one by one as Sen _ scoreij。
The first group of attribute feature-emotion word pairs comprise the 6 th attribute feature 'running speed' in the attribute feature set, and the emotion polarity of the adjective 'Kadun' in the emotion words is marked as beta61The emotion words do not contain the adverbs for modifying the emotion of the adjectives, but contain a negative word, the number f of the negative words is an odd number, and the score mu of the negative word corresponding to the emotion words is-1. Finally, the calculation mode of the basic score of the set of attribute feature-emotion word pairs is Sen _ score61＝(-1)×β61. The second group of attribute features, namely the 2 nd attribute 'appearance' in the attribute set contained in the emotional word pair, scores the emotional polarity of the adjective 'good looking' in the emotional word as beta21The emotion word does not include a process for modifying the emotion of the adjectiveDegree adverb, negation, and adjective emotion polarity score beta in emotional words21Is the base score Sen _ score of the set of attribute feature-emotion word pairs21. The third group of attribute features, namely the 2 nd attribute 'appearance' in the attribute set contained in the emotional word pair, scores the emotional polarity of the adjective 'good looking' in the emotional word as beta22The degree adverb "too" in the emotional word, which does not contain a negative word but contains a modified adjective, corresponds to the coefficient phi of the degree levelDegreeThen the basic score calculation mode of the set of attribute feature-emotion word pairs is Sen _ score21＝φDegree×β22。
With "appearance" properties, i.e. F2For example, the emotion polarity base scores of all emotion words for that attribute are summed and divided by the associated attribute F in the evaluation dataset2Total number of occurrences of emotional words n (F)2) To obtain the sentiment value Sen _ score (F) of the 'appearance' attribute2)。
After obtaining the emotion values of all attributes, the emotion value Sen _ score (F) of a single attributei) Multiplying the attribute by a weight coefficient alpha corresponding to the attributeiAnd then summing to obtain the total commodity score of the mobile phone product based on the attribute weight.
5. Analysis result display
The displayed analysis result comprises descriptive statistics such as data set size, word frequency statistics and evaluation text word cloud pictures generated by the module III and commodity attribute scores, commodity attribute score radar maps, commodity total scores based on attribute weights and system operation conditions obtained by calculation of the module V.
Claims (7)
1. The commodity evaluation emotion analysis system based on attribute characteristics is characterized by comprising the following modules:
a user interaction module: the system comprises a network address used for authenticating the identity of a user and receiving a target commodity input by the user;
an evaluation statement acquisition module: capturing commodity details and consumer evaluation data from a website of a target commodity by using a crawler technology;
a data preprocessing module: the evaluation statement acquisition module is used for acquiring evaluation statements of the user;
an extraction module: the system is used for extracting the attribute characteristics and the attribute weight of the commodity from the preprocessed data, and obtaining emotional words matched with the attribute characteristics from the consumer evaluation data through semantic relation calculation, word segmentation and part-of-speech tagging methods so as to obtain attribute characteristic-emotional word pairs;
an algorithm processing module: the system is used for inputting the extracted attribute feature-emotion word pair into an algorithm model for computational analysis, realizing an emotion analysis algorithm based on the attribute feature and outputting a commodity evaluation emotion quantitative value;
and a result display module: used for displaying the analysis result.
2. The system for analyzing commodity evaluation emotion based on attribute characteristics of claim 1, wherein the preprocessing operation of the data preprocessing module includes: word segmentation, filtering, part of speech tagging and word frequency statistics.
3. The system for analyzing commodity evaluation emotion based on attribute features as claimed in claim 2, wherein the extracting module extracts the attribute features of the commodity including explicit attribute features and implicit attribute features, and the specific step of extracting the attribute features includes:
determining the explicit attribute characteristics of the commodity according to the commodity details and the consumer evaluation information;
screening out all nouns and nominal phrases in the comment set as a candidate word set through the statistical word frequency;
through point mutual information calculation, a noun or a nominal phrase with high mutual information value with the explicit attribute characteristics in the candidate word set is identified and is used as the implicit attribute characteristics.
4. The system for analyzing commodity evaluation emotion based on attribute characteristics of claim 3, wherein the formula of the point-to-point information calculation is as follows:
wherein, PMI (F)iPh) as an explicit Attribute feature FiMutual information value with noun or noun phrase ph, ph being noun or noun phrase in comment set, p (F)iPh) explicit Attribute feature F for comment setsiProbability of co-occurrence with noun or noun phrase ph, p (F)i) Aggregating explicit attribute features for comments FiThe probability of occurrence, p (ph), is the probability of occurrence of the candidate attribute word in the comment set.
5. The system for analyzing commodity evaluation emotion based on attribute characteristics as claimed in claim 2, wherein the extraction module obtains the attribute weight of the commodity according to the word frequency statistics result.
6. The system for analyzing commodity evaluation emotion based on attribute feature of claim 1, wherein the manner in which the algorithm processing module implements the emotion analysis algorithm based on attribute feature is as follows:
single set of attribute feature-emotion word pairs<Feature(i),Opinion(j)>Sentiment polarity base score of Sen _ scoreijIs defined as:
then the emotion scoring algorithm for the individual attribute features of a certain commodity is as follows:
the sentiment score of a single good is formulated as follows:
wherein m represents the number of attribute features contained in a single commodity, n is the number of emotion words appearing in the attribute features,
for the degree of adverb intensity level, phi, occurring before each emotional wordDegreeCoefficient values corresponding to the strength levels of the adverbs with corresponding degrees, f is the number of negative words before the emotional word, mu is the score value of the negative word of the emotional word, class1-class4 are the strength levels of the adverbs with four divided degrees, and alphaiIs the weight coefficient of the attribute feature, i is the serial number of the attribute feature, i belongs to {1,2,3, … …, m }, betaijSentiment polarity base score, beta, for the jth sentiment word of the ith attribute featureijThe value range is [ -1,1 [ ]]Wherein positive and negative correspond to positive and negative emotions, respectively.
7. The system for analyzing commodity evaluation emotion based on attribute characteristics of claim 1, wherein the analysis result presented by the result presentation module includes: the method comprises the following steps of commodity attribute scoring, radar chart scoring of each attribute of the commodity, total commodity scoring based on attribute weight and system running conditions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910916447.6A CN110706028A (en) | 2019-09-26 | 2019-09-26 | Commodity evaluation emotion analysis system based on attribute characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910916447.6A CN110706028A (en) | 2019-09-26 | 2019-09-26 | Commodity evaluation emotion analysis system based on attribute characteristics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110706028A true CN110706028A (en) | 2020-01-17 |
Family
ID=69196465
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910916447.6A Pending CN110706028A (en) | 2019-09-26 | 2019-09-26 | Commodity evaluation emotion analysis system based on attribute characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110706028A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310455A (en) * | 2020-02-11 | 2020-06-19 | 安徽理工大学 | New emotion word polarity calculation method for online shopping comments |
| CN112418919A (en) * | 2020-11-10 | 2021-02-26 | 江南大学 | Hairy crab quality evaluation method based on big data screening and instrument analysis |
| CN112464762A (en) * | 2020-11-16 | 2021-03-09 | 中国科学院合肥物质科学研究院 | Agricultural product screening system and method based on image processing |
| CN112785335A (en) * | 2021-01-21 | 2021-05-11 | 安徽商信政通信息技术股份有限公司 | Data processing method and system for electronic government affair performance assessment system |
| WO2021243956A1 (en) * | 2020-06-05 | 2021-12-09 | 浙江工商大学 | Method for performing fine-grained text sentiment analysis on basis of degree of user dissatisfaction |
| CN117114746A (en) * | 2023-08-24 | 2023-11-24 | 哈尔滨工业大学 | Method for predicting emotion of consumer by sudden public health event |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130091117A1 (en) * | 2011-09-30 | 2013-04-11 | Metavana, Inc. | Sentiment Analysis From Social Media Content |
| CN103455562A (en) * | 2013-08-13 | 2013-12-18 | 西安建筑科技大学 | Text orientation analysis method and product review orientation discriminator on basis of same |
| CN103678564A (en) * | 2013-12-09 | 2014-03-26 | 国家计算机网络与信息安全管理中心 | Internet product research system based on data mining |
| CN104699766A (en) * | 2015-02-15 | 2015-06-10 | 浙江理工大学 | Implicit attribute mining method integrating word correlation and context deduction |
| CN106127507A (en) * | 2016-06-13 | 2016-11-16 | 四川长虹电器股份有限公司 | A kind of commodity the analysis of public opinion method and system based on user's evaluation information |
| CN110175325A (en) * | 2019-04-26 | 2019-08-27 | 南京邮电大学 | The comment and analysis method and Visual Intelligent Interface Model of word-based vector sum syntactic feature |
-
2019
- 2019-09-26 CN CN201910916447.6A patent/CN110706028A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130091117A1 (en) * | 2011-09-30 | 2013-04-11 | Metavana, Inc. | Sentiment Analysis From Social Media Content |
| CN103455562A (en) * | 2013-08-13 | 2013-12-18 | 西安建筑科技大学 | Text orientation analysis method and product review orientation discriminator on basis of same |
| CN103678564A (en) * | 2013-12-09 | 2014-03-26 | 国家计算机网络与信息安全管理中心 | Internet product research system based on data mining |
| CN104699766A (en) * | 2015-02-15 | 2015-06-10 | 浙江理工大学 | Implicit attribute mining method integrating word correlation and context deduction |
| CN106127507A (en) * | 2016-06-13 | 2016-11-16 | 四川长虹电器股份有限公司 | A kind of commodity the analysis of public opinion method and system based on user's evaluation information |
| CN110175325A (en) * | 2019-04-26 | 2019-08-27 | 南京邮电大学 | The comment and analysis method and Visual Intelligent Interface Model of word-based vector sum syntactic feature |
Non-Patent Citations (1)
| Title |
|---|
| 许云峰 等著: "《大数据技术及行业应用》", 31 August 2016, 北京邮电大学出版 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310455A (en) * | 2020-02-11 | 2020-06-19 | 安徽理工大学 | New emotion word polarity calculation method for online shopping comments |
| WO2021243956A1 (en) * | 2020-06-05 | 2021-12-09 | 浙江工商大学 | Method for performing fine-grained text sentiment analysis on basis of degree of user dissatisfaction |
| US11748565B2 (en) | 2020-06-05 | 2023-09-05 | Zhejiang Gongshang University | Method for analyzing fine-grained text sentiment based on users' harshness |
| CN112418919A (en) * | 2020-11-10 | 2021-02-26 | 江南大学 | Hairy crab quality evaluation method based on big data screening and instrument analysis |
| CN112464762A (en) * | 2020-11-16 | 2021-03-09 | 中国科学院合肥物质科学研究院 | Agricultural product screening system and method based on image processing |
| CN112785335A (en) * | 2021-01-21 | 2021-05-11 | 安徽商信政通信息技术股份有限公司 | Data processing method and system for electronic government affair performance assessment system |
| CN117114746A (en) * | 2023-08-24 | 2023-11-24 | 哈尔滨工业大学 | Method for predicting emotion of consumer by sudden public health event |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10748164B2 (en) | Analyzing sentiment in product reviews | |
| CN110008311B (en) | Product information safety risk monitoring method based on semantic analysis | |
| Kumar et al. | Sentiment analysis of multimodal twitter data | |
| CN108491377B (en) | E-commerce product comprehensive scoring method based on multi-dimensional information fusion | |
| CN109933664B (en) | Fine-grained emotion analysis improvement method based on emotion word embedding | |
| CN111767741B (en) | Text emotion analysis method based on deep learning and TFIDF algorithm | |
| CN110706028A (en) | Commodity evaluation emotion analysis system based on attribute characteristics | |
| CN103778214B (en) | A kind of item property clustering method based on user comment | |
| CN103870973B (en) | Information push, searching method and the device of keyword extraction based on electronic information | |
| CN108388660B (en) | Improved E-commerce product pain point analysis method | |
| US10366117B2 (en) | Computer-implemented systems and methods for taxonomy development | |
| CN108319734A (en) | A kind of product feature structure tree method for auto constructing based on linear combiner | |
| CN112015721A (en) | E-commerce platform storage database optimization method based on big data | |
| Yang et al. | A decision method for online purchases considering dynamic information preference based on sentiment orientation classification and discrete DIFWA operators | |
| CN111242679A (en) | Sales forecasting method based on product review viewpoint mining | |
| CN102929860B (en) | Chinese clause emotion polarity distinguishing method based on context | |
| Shen et al. | A voice of the customer real-time strategy: An integrated quality function deployment approach | |
| CN112182145A (en) | Text similarity determination method, device, equipment and storage medium | |
| Chen et al. | Research on credit evaluation model of online store based on SnowNLP | |
| Leopairote et al. | Software quality in use characteristic mining from customer reviews | |
| CN117764669A (en) | Article recommendation method, device, equipment, medium and product | |
| CN113010639A (en) | Commodity analysis method and device based on E-commerce platform | |
| CN107291686B (en) | Method and system for identifying emotion identification | |
| Yang et al. | Research on the Sentiment analysis of customer reviews based on the ontology of phone | |
| KR20220118703A (en) | Machine Learning based Online Shopping Review Sentiment Prediction System and Method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200117 |
|
| RJ01 | Rejection of invention patent application after publication |