CN110442728A - Sentiment dictionary construction method based on word2vec automobile product field - Google Patents
Sentiment dictionary construction method based on word2vec automobile product field Download PDFInfo
- Publication number
- CN110442728A CN110442728A CN201910580589.XA CN201910580589A CN110442728A CN 110442728 A CN110442728 A CN 110442728A CN 201910580589 A CN201910580589 A CN 201910580589A CN 110442728 A CN110442728 A CN 110442728A
- Authority
- CN
- China
- Prior art keywords
- data
- sentiment dictionary
- sentiment
- word2vec
- dictionary
- 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
- 238000010276 construction Methods 0.000 title claims abstract description 14
- 230000008451 emotion Effects 0.000 claims abstract description 34
- 239000000047 product Substances 0.000 claims abstract description 20
- 238000012549 training Methods 0.000 claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 14
- 238000005516 engineering process Methods 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 8
- 239000013589 supplement Substances 0.000 claims abstract description 4
- 238000004140 cleaning Methods 0.000 claims abstract description 3
- 238000012216 screening Methods 0.000 claims description 7
- 238000005259 measurement Methods 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 4
- 230000008676 import Effects 0.000 claims description 2
- 238000005201 scrubbing Methods 0.000 claims description 2
- 238000004458 analytical method Methods 0.000 abstract description 9
- 238000004364 calculation method Methods 0.000 abstract description 4
- 238000006243 chemical reaction Methods 0.000 abstract description 2
- 230000008901 benefit Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 206010028916 Neologism Diseases 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- 238000012552 review Methods 0.000 description 3
- 230000008909 emotion recognition Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 241001178520 Stomatepia mongo Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 235000013372 meat Nutrition 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000003442 weekly effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/374—Thesaurus
Abstract
The sentiment dictionary construction method in the invention discloses a kind of automobile product field based on word2vec, it include: to take off to take automotive vertical class website user's comment data using python orientation crawler technology, parsing obtains the structural data based on website-vehicle-public praise and stores into postgresql database；Data cleansing, cleaning, conversion between simplified and traditional Chinese of processing, noise data including shortage of data value etc. are carried out to public praise data in database；Selected part text carries out Emotion tagging as training set, wherein 1 is forward direction, 0 is negative sense；Model training is carried out using word2vec, high-volume text data is imported into training pattern and carries out similarity calculation；Initial sentiment dictionary is further expanded based on existing sentiment dictionary；Output text obtains sentiment dictionary, and finally carries out artificial reversed emotion word supplement.The present invention solves the problems, such as that the sentiment dictionary constructed based on artificial and knowledge based library method analyzes inaccuracy when handling automotive field sentiment analysis.
Description
Technical field
The present invention relates to identification field more particularly to a kind of sentiment dictionary structures in the automobile product field based on word2vec
Construction method.
Background technique
On 2 28th, 2019, China Internet Network Information Center (CNNIC) 43rd " China Internet network of publication in Beijing
State of development statistical report "[1].By in December, 2018, China's netizen's scale is up to 8.29 hundred million, and whole year increases netizen 56,530,000 newly, mutually
Popularity rate of networking is 59.6%, promotes 3.8 percentage points compared with the end of the year 2017.With the fast development of netizen's scale, data also with
Blowout generate and rapidly accumulate." science and technology prophesy Supreme Being " triumphant benefit of Kevin says that, to the year two thousand fifty, global metadata total amount is up to
The astronomical magnitude of 1000000 ZB.
With the popularity of the internet, user can express oneself emotion and attitude by internet platform whenever and wherever possible.It is logical
It crosses and mining analysis is carried out to these comment texts with emotional color, can effectively help businessman's focused user most concerned about asking
Topic and item property understand user to product customer satisfaction, to targetedly continue previous generation in next-generation iteration
Product advantage improves the problem of user proposes, promotes the market competitiveness.Therefore, the use for lying in product review behind is excavated
Family Sentiment orientation can help enterprise discovery consumer to concentrate the defect complained, more efficiently to cater to the need of consumer
It asks[2]。
In user feeling trend analysis, sentiment dictionary is indispensable as most important analysis tool.But at present in
Literary sentiment dictionary is relatively fewer, sentiment dictionary and incompatible in each field, and versatility is lower, and same emotion word is in different context
In also have different emotions tendency, as " chassis high " and emotion expressed by " oil consumption height " are completely contradicted.
To help automobile vendor to fully understand Related product attribute client's Sentiment orientation, the present invention is towards automotive vertical field
Website platform establish automobile product field sentiment dictionary in conjunction with text mining, data processing and word2vec technology, be
The accurate control customer priorities of automobile vendor lay the foundation.
In the implementation of the present invention, discovery at least has the following disadvantages in the prior art and deficiency by inventor:
1) manual sorting sentiment dictionary, low efficiency；
2) sentiment dictionary, coverage area are created with General Inquirer (GI), SentiWordNet semiautomatic fashion
Limited, field adaptability and reliability are poor[3][4], and a large amount of artificial mark work is needed, efficiency is lower, vulnerable to subjectivity
Property influence, intensity mark fine granularity also cannot be guaranteed with accuracy；
3) expand sentiment dictionary using semantic knowledge-base, such as WordNet, HowNet[5][6], semantic knowledge-base is excessively relied on,
Limited coverage area, field applicability are poor；
4) new term and mutation vocabulary continue to bring out, the continuing to bring out of various network neologisms cause the lag of dictionary with
And dictionary itself quality the problems such as, make sentiment dictionary that can not be matched to effectively emotion word, cause analysis fail, therefore need pair
Dictionary carries out real-time update.
Bibliography
[1] CNNIC, which issues the 43rd " China Internet network state of development statistical report " [J] and nets, believes civil-military inosculation, 2019
(02):37-38.
[2]Guo W,Liang R Y,Wang L,et al.Exploring sustained participation in
firm-hosted communities in China:the effects of social capital and active
degree[J].Behaviour&Information Technology,2017,36(3):223-242.
[3] Wang Ke, Xia Rui sentiment dictionary method for auto constructing summarize [J] and automate journal, 2016,42 (04): 495-
511.
[4] county Xie Song, Liu Bo, Wang Ting application semantics relationship construct sentiment dictionary [J] National University of Defense technology journal automatically,
2014,36(03):111-115.
[5] Kim S M, Hovy E.Determining the Sentiment of Opinions [C]
.Proceedings of the 20th International Conference on Computational
Linguistics.Association for Computational Linguistics,2004.
[6] Hassan A, Radev D.Identifying Text Polarity Using Random Walks [C]
.Proceedings of the 48th
Annual Meeting of the Association for Computational
Linguistics.Association for Computational Linguistics,2010:395-403.
Summary of the invention
Internet automotive vertical of the present invention website, in conjunction with text mining, data processing and word2vec technology, building
A kind of sentiment dictionary construction method in the automobile product field based on word2vec, solves artificial constructed sentiment dictionary efficiency
It is low, excessively rely on WordNet, HowNet semantic knowledge-base, coverage area is narrow, field is poor for applicability and emotion new word identification
The problems such as, it is described below:
A kind of sentiment dictionary construction method in the automobile product field based on word2vec, the method includes following steps
It is rapid:
It is taken off using python orientation crawler technology and takes the data such as automotive vertical class website user comment, parsing is obtained with net
Stand-vehicle-public praise based on structural data and store into database；
Data scrubbing, cleaning, the letter of processing, noise data including shortage of data value are carried out to public praise data in database
Numerous conversion etc.；
Selected part text carries out Emotion tagging as training set；Model training is carried out using word2vec, it will high-volume
Text data imports training pattern and carries out similarity judgement, obtains initial sentiment dictionary based on output text；
In conjunction with existing sentiment dictionary to identification or the lesser word progress Semantic Similarity Measurement of similarity is failed, to emotion
Dictionary is expanded；
Artificial screening is carried out to above-mentioned gained sentiment dictionary and carries out reversed emotion word supplement, output obtains final emotion word
Allusion quotation.
Wherein, the crawl initial data specifically:
1) model data, comprising: vehicle classification, vehicle brand and vehicle price range in Vertical Website；
2) public praise data, comprising: user is directed to the evaluation content and correlation based on bus attribute that different automobile types are made
Evaluation time.
Further, the data grab method specifically:
MRQ distributed data acquisition frame is constructed, the user comment module of multiple automotive vertical classes website is grabbed, with " net
Stand-vehicle-public praise " tree structure deposit postgresql database in, and increment crawl periodically is carried out to it.
Wherein, the data training set acquisition methods specifically:
Appropriate weigh is assigned according to comment time, vehicle price range with " most more unsatisfied " to " a bit most satisfied "
Weight is trained selecting at random for collection based on this, and positive negative emotion sentence is labeled.
Further, the similarity calculating method specifically:
Similarity judgement, correlation formula are carried out based on term vector cosine size are as follows:
Wherein, the sentiment dictionary extending method specifically:
Dalian University of Technology's Chinese emotion vocabulary ontology library, Hownet sentiment dictionary and the thesaurus of Harbin Institute of Technology are carried out
Duplicate removal and arrangement；
The word trained is filtered, filters out and fails identification or the lesser word of similarity, it is gone to reform together
Emotion benchmark word dictionary after reason carries out Semantic Similarity Measurement, finally determines its Sentiment orientation；
Artificial further screening is carried out to above-mentioned sentiment dictionary, and is reversely expanded, final sentiment dictionary is exported.
The beneficial effect of the technical scheme provided by the present invention is that:
1, the inefficiency of the invention for solving manual sorting sentiment dictionary, General Inquirer (GI),
SentiWordNet semiautomatic fashion creates sentiment dictionary field adaptability and reliability is poor and needs largely manually mark
The lower problem of work, efficiency；
2, the present invention solves WordNet, HowNet and excessively relies on semantic knowledge-base, and various network neologisms continue to bring out
Caused by dictionary lag the problem of；
3, the present invention passes through comprehensive Dalian University of Technology's Chinese emotion vocabulary ontology library and Hownet sentiment dictionary
(HOWNET), dictionary is expanded based on semantic similarity, it is not comprehensive solves the constructed sentiment dictionary covering of conventional method
The problem of；
4, the present invention was solved based on the constructed sentiment dictionary of semantic knowledge-base in the processing automotive field sentiment analysis time-division
The problem for analysing inaccuracy facilitates Market Analyst more accurately to hold the Sentiment orientation of the bus attribute of user's description.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the sentiment dictionary construction method in automobile product field based on word2vec；
Fig. 2 is the schematic diagram of data store organisation；
Fig. 3 is to carry out emotion based on word2vec to differentiate schematic diagram；
Fig. 4 is that sentiment dictionary expands schematic diagram.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, embodiment of the present invention is made below further
Ground detailed description.
Embodiment 1
The sentiment dictionary construction method in the embodiment of the invention provides a kind of automobile product field based on word2vec, ginseng
See Fig. 1, method includes the following steps:
101: utilizing MRQ (distributed task scheduling queue of the Python based on Redis, Mongo and gevent) distributed data
Collection mechanism grabs the user comment module of multiple automotive vertical classes website, and with the tree structure of " website-vehicle-public praise "
It is stored in postgresql database；
102: extracting the public praise data in database, public praise data include " a bit most satisfied ", " most unsatisfied one altogether
Mouth is chosen in point ", " space ", " power ", " manipulation ", " oil consumption ", " comfort ", " appearance ", " interior trim ", " cost performance " ten part
In landmark data " a bit most satisfied " and " most more unsatisfied ", pre-process to data are chosen: rejecting abnormalities are commented on, with
Punctuation mark is that cut point is cut, and change long sentence is short sentence；
103: randomly selecting comment data as training sample set, wherein institute's sample drawn need to be useful comment: i.e. simultaneously
Complete sentence comprising emotion word and attribute word.Then it is manually marked, wherein front comment is designated as 1, negative reviews
It is designated as 0；
104: training sample set being segmented, stop words is gone to handle, using text after processing as the input of word2vec
File simultaneously specifies suitable training parameter, carries out the training of Chinese term vector, and carries out similarity calculation；
105: to Dalian University of Technology's Chinese emotion vocabulary ontology library, Hownet sentiment dictionary (HOWNET) and Harbin Institute of Technology
Thesaurus carries out duplicate removal and arrangement, forms emotion benchmark word dictionary.The word trained is filtered, filters out and fails to know
Its emotion benchmark word dictionary after arranging with duplicate removal is carried out Semantic Similarity Measurement, finally by the other or lesser word of similarity
Determine its Sentiment orientation；
106: artificial screening training result, and carry out the supplement of reversed emotion word.
In conclusion 101- step 106 realizes the feelings to automobile product field to the embodiment of the present invention through the above steps
The building for feeling dictionary improves accuracy rate when carrying out sentiment analysis to automotive field.
Embodiment 2
The scheme in embodiment 1 is further introduced below with reference to specific calculation formula, example, it is as detailed below
Description:
201: the acceptance of the users data of multiple automotive vertical websites being obtained by data grabber and are stored into database；
Wherein, the step 201 specifically:
1) MRQ distributed data acquisition mechanism is based on by python language and is directed to automotive vertical class to be crawled website
Programming is carried out, grabs information needed webpage source code, and parsed to obtain specifically to webpage source code based on regular expression
Comment data under vehicle model information and the vehicle.
The data of above-mentioned crawl include web site url, webpage source code, user information, vehicle model information, time of posting etc., wherein
Predominantly acceptance of the users data.Acceptance of the users data are classified again are as follows: " a bit most satisfied ", " most more unsatisfied ", " sky
Between ", " power ", " manipulation ", " appearance ", " interior trim ", " oil consumption ", " comfort ", " cost performance "；
2) data of crawl are stored in local data base with the tree structure of " website-vehicle-public praise ", such as Fig. 2 institute
Show；
3) increment crawl is carried out to above-mentioned data weekly, with the emerging bus attribute of correspondence and network words.
202: experimental data prepares
1) by sql sentence reading-port landmark data from database, public praise data is divided by type then, are respectively obtained
" a bit most satisfied " and " most more unsatisfied "；
2) long comment text is split using punctuate as cut-point, obtains short comment text；
3) short comment text is further screened, rejects garbage and abnormal comment, reduces noise.
203: comment text is randomly selected according to time, vehicle grade and vehicle price range；
1) influence of the other factors to experimental result is reduced for minimum limit, to short comment text obtained by step 202, according to
Model year, vehicle grade and vehicle price range are divided, and are randomly selected comment text according to appropriate weight, are guaranteed comment
Data coverage is comprehensive, increases universality.Wherein front comment is 200 total, and negative reviews are 400 total；
2) institute's extracting comment text is manually marked, it, need to be right by three when mark wherein 1 is negative sense for forward direction, 0
Automobile knowledge has the classmate of basic understanding independently to mark, and selects in three annotation results maximum value as annotation results.
204: marking pretreatment and the model training of text
1) it is loaded into jieba participle, marked text is segmented, stop words dictionary is then subsequently loaded into, is gone unless retrieving
Word, as " ", Keyword Density is improved, memory space is saved and improves search efficiency；
2) text after processing as the input file of word2vec and is specified into suitable training parameter, passes through deep learning
Thought carry out the training of Chinese term vector, content of text is mapped in higher-dimension vector row space, k dimensional vector is formed, then
Similarity judgement is carried out by calculating term vector cosine size, the expansion and the emotion recognition of neologisms of near synonym are realized with this,
Such as two k dimensional vector m (x11,x12,x13,…,x1k),n(x21,x22,x23,…,x2k), cosine calculation formula is as follows；
205: the extension of emotion dictionary
1) due to more wrong words, network words, phonogram etc. in comment, obtained dictionary presence covering is incomplete can
Can, therefore step 204 sentiment dictionary obtained need to be expanded, i.e., it will not be appeared in context in existing sentiment dictionary
Word emotion recognition and expand.
First to Dalian University of Technology's Chinese emotion vocabulary ontology library, Hownet sentiment dictionary (HOWNET) and Harbin Institute of Technology
Thesaurus carries out duplicate removal and arrangement, obtains completely new emotion benchmark word dictionary.
2) word that step 204 trains is filtered, filters out and fail identification or the lesser word of similarity, by it
Semantic Similarity Measurement, which is carried out, with the emotion benchmark word dictionary after duplicate removal finally determines its Sentiment orientation.
206: reversed emotion word expansion is carried out to sentiment dictionary
1) training result is subjected to artificial screening first, removal does not meet the type of error of context；
2) expansion that reversed emotion word is carried out to result after screening, after above-mentioned dictionary is formed, there are part attributes only to include
Positive or negative emotion, such as " starting meat ", " chassis is low ", need to be expanded at this time, artificial addition " starting is fast ", " chassis is high "
Deng for not only including negative in attribute but also then no longer being handled comprising positive sentence.
It is to sum up told, the embodiment of the present invention can obtain the sentiment dictionary in automobile product field, built feelings through the above steps
The high accuracy and availability for feeling dictionary improve accuracy rate when carrying out sentiment analysis to automotive field, facilitate market analysis
Personnel more accurately hold the Sentiment orientation of bus attribute described in user.
The embodiment of the present invention to the model of each device in addition to doing specified otherwise, the model of other devices with no restrictions,
As long as the device of above-mentioned function can be completed.
It will be appreciated by those skilled in the art that attached drawing is the schematic diagram of a preferred embodiment, the embodiments of the present invention
Serial number is for illustration only, does not represent the advantages or disadvantages of the embodiments.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (6)
1. a kind of sentiment dictionary construction method in the automobile product field based on word2vec, which is characterized in that the method packet
Include following steps:
It is taken off using python orientation crawler technology and takes the data such as automotive vertical class website user comment, parsing is obtained with website-vehicle
Structural data and storage based on type-public praise is into database；
Data scrubbing, the cleaning of processing, noise data including shortage of data value, simplified and traditional turn are carried out to public praise data in database
It changes；
Selected part text carries out Emotion tagging as training set；Model training is carried out using word2vec, by high-volume text
Data import training pattern and carry out similarity judgement, obtain initial sentiment dictionary based on output text；
In conjunction with existing sentiment dictionary to identification or the lesser word progress Semantic Similarity Measurement of similarity is failed, to sentiment dictionary
Expanded；
Artificial screening is carried out to above-mentioned gained sentiment dictionary and carries out reversed emotion word supplement, output obtains final sentiment dictionary.
2. a kind of sentiment dictionary construction method in automobile product field based on word2vec according to claim 1,
It is characterized in that, the crawl initial data specifically:
1) model data, comprising: vehicle classification, vehicle brand and vehicle price range in Vertical Website；
2) public praise data, comprising: user is directed to the evaluation content and relevant evaluation based on bus attribute that different automobile types are made
Time.
3. a kind of sentiment dictionary construction method in automobile product field based on word2vec according to claim 1,
It is characterized in that, the data grab method specifically:
MRQ distributed data acquisition frame is constructed, the user comment module of multiple automotive vertical classes website is grabbed, with " website-
In the tree structure deposit postgresql database of vehicle-public praise ", and increment crawl periodically is carried out to it.
4. a kind of sentiment dictionary construction method in automobile product field based on word2vec according to claim 1,
It is characterized in that, the data training set acquisition methods specifically:
Appropriate weight is assigned according to comment time, vehicle price range with " most more unsatisfied " to " a bit most satisfied ",
It is trained selecting at random for collection based on this, and positive negative emotion sentence is labeled.
5. a kind of sentiment dictionary construction method in automobile product field based on word2vec according to claim 1,
It is characterized in that, the similarity calculating method specifically:
Similarity judgement, correlation formula are carried out based on term vector cosine size are as follows:
6. a kind of sentiment dictionary construction method in automobile product field based on word2vec according to claim 1,
It is characterized in that, the sentiment dictionary extending method specifically:
Duplicate removal is carried out to Dalian University of Technology's Chinese emotion vocabulary ontology library, Hownet sentiment dictionary and the thesaurus of Harbin Institute of Technology
And arrangement；
The word trained is filtered, filters out and fails identification or the lesser word of similarity, after it is arranged with duplicate removal
Emotion benchmark word dictionary carry out Semantic Similarity Measurement, finally determine its Sentiment orientation；
Artificial further screening is carried out to above-mentioned sentiment dictionary, and is reversely expanded, final sentiment dictionary is exported.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910580589.XA CN110442728A (en) | 2019-06-28 | 2019-06-28 | Sentiment dictionary construction method based on word2vec automobile product field |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910580589.XA CN110442728A (en) | 2019-06-28 | 2019-06-28 | Sentiment dictionary construction method based on word2vec automobile product field |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110442728A true CN110442728A (en) | 2019-11-12 |
Family
ID=68428826
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910580589.XA Pending CN110442728A (en) | 2019-06-28 | 2019-06-28 | Sentiment dictionary construction method based on word2vec automobile product field |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110442728A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111144929A (en) * | 2019-12-04 | 2020-05-12 | 天津大学 | Comment object and word combined extraction method for automobile industry user generated content |
| CN111612015A (en) * | 2020-05-26 | 2020-09-01 | 创新奇智(西安)科技有限公司 | Vehicle identification method and device and electronic equipment |
| CN111724196A (en) * | 2020-05-14 | 2020-09-29 | 天津大学 | Method for improving quality of automobile product based on user experience |
| CN111931497A (en) * | 2020-07-16 | 2020-11-13 | 中国汽车技术研究中心有限公司 | Optimization method for language of questionnaire for automobile consumer |
| CN112016964A (en) * | 2020-08-27 | 2020-12-01 | 李忠耘 | Second-hand vehicle dynamic pricing method and device, electronic equipment and storage medium |
| CN113761882A (en) * | 2020-06-08 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Dictionary construction method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130103623A1 (en) * | 2011-10-21 | 2013-04-25 | Educational Testing Service | Computer-Implemented Systems and Methods for Detection of Sentiment in Writing |
| CN107704556A (en) * | 2017-09-28 | 2018-02-16 | 北京车慧互动广告有限公司 | A kind of sentiment analysis method and system in automobile industry subdivision field |
| CN109408809A (en) * | 2018-09-25 | 2019-03-01 | 天津大学 | A kind of sentiment analysis method for automobile product comment based on term vector |
-
2019
- 2019-06-28 CN CN201910580589.XA patent/CN110442728A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130103623A1 (en) * | 2011-10-21 | 2013-04-25 | Educational Testing Service | Computer-Implemented Systems and Methods for Detection of Sentiment in Writing |
| CN107704556A (en) * | 2017-09-28 | 2018-02-16 | 北京车慧互动广告有限公司 | A kind of sentiment analysis method and system in automobile industry subdivision field |
| CN109408809A (en) * | 2018-09-25 | 2019-03-01 | 天津大学 | A kind of sentiment analysis method for automobile product comment based on term vector |
Non-Patent Citations (1)
| Title |
|---|
| 陆峰: "基于word2vec扩充情感词典的商品评论倾向分析", 《电脑知识与技术》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111144929A (en) * | 2019-12-04 | 2020-05-12 | 天津大学 | Comment object and word combined extraction method for automobile industry user generated content |
| CN111724196A (en) * | 2020-05-14 | 2020-09-29 | 天津大学 | Method for improving quality of automobile product based on user experience |
| CN111612015A (en) * | 2020-05-26 | 2020-09-01 | 创新奇智(西安)科技有限公司 | Vehicle identification method and device and electronic equipment |
| CN111612015B (en) * | 2020-05-26 | 2023-10-31 | 创新奇智(西安)科技有限公司 | Vehicle identification method and device and electronic equipment |
| CN113761882A (en) * | 2020-06-08 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Dictionary construction method and device |
| CN111931497A (en) * | 2020-07-16 | 2020-11-13 | 中国汽车技术研究中心有限公司 | Optimization method for language of questionnaire for automobile consumer |
| CN112016964A (en) * | 2020-08-27 | 2020-12-01 | 李忠耘 | Second-hand vehicle dynamic pricing method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111737495B (en) | Middle-high-end talent intelligent recommendation system and method based on domain self-classification | |
| CN110633373B (en) | Automobile public opinion analysis method based on knowledge graph and deep learning | |
| CN111914096B (en) | Public opinion knowledge graph-based public transportation passenger satisfaction evaluation method and system | |
| CN108573411B (en) | Mixed recommendation method based on deep emotion analysis and multi-source recommendation view fusion of user comments | |
| CN110633409B (en) | Automobile news event extraction method integrating rules and deep learning | |
| CN110442728A (en) | Sentiment dictionary construction method based on word2vec automobile product field | |
| CN102831184B (en) | According to the method and system text description of social event being predicted to social affection | |
| CN110555568B (en) | Road traffic running state real-time perception method based on social network information | |
| CN109684440A (en) | Address method for measuring similarity based on level mark | |
| CN105528437B (en) | A kind of question answering system construction method extracted based on structured text knowledge | |
| CN109829166B (en) | People and host customer opinion mining method based on character-level convolutional neural network | |
| CN106649260A (en) | Product feature structure tree construction method based on comment text mining | |
| CN109344187B (en) | Structured processing system for judicial judgment case information | |
| CN105512687A (en) | Emotion classification model training and textual emotion polarity analysis method and system | |
| CN107885793A (en) | A kind of hot microblog topic analyzing and predicting method and system | |
| CN109408809A (en) | A kind of sentiment analysis method for automobile product comment based on term vector | |
| CN107977798A (en) | A kind of risk evaluating method of e-commerce product quality | |
| CN110807091A (en) | Hotel intelligent question-answer recommendation and decision support analysis method and system | |
| CN103838833A (en) | Full-text retrieval system based on semantic analysis of relevant words | |
| CN109726274A (en) | Problem generation method, device and storage medium | |
| CN111353044B (en) | Comment-based emotion analysis method and system | |
| CN112183056A (en) | Context-dependent multi-classification emotion analysis method and system based on CNN-BilSTM framework | |
| CN107967290A (en) | A kind of knowledge mapping network establishing method and system, medium based on magnanimity scientific research data | |
| CN109325860A (en) | Network public-opinion detection method and system for overseas investment Risk-warning | |
| CN115796181A (en) | Text relation extraction method for chemical field |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191112 |