CN109213861A - In conjunction with the tourism evaluation sensibility classification method of At_GRU neural network and sentiment dictionary - Google Patents
In conjunction with the tourism evaluation sensibility classification method of At_GRU neural network and sentiment dictionary Download PDFInfo
- Publication number
- CN109213861A CN109213861A CN201810862476.4A CN201810862476A CN109213861A CN 109213861 A CN109213861 A CN 109213861A CN 201810862476 A CN201810862476 A CN 201810862476A CN 109213861 A CN109213861 A CN 109213861A
- Authority
- CN
- China
- Prior art keywords
- tourism
- word
- evaluation
- sentiment dictionary
- neural network
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Machine Translation (AREA)
Abstract
The present invention relates to the tourism evaluation sensibility classification methods of a kind of combination At_GRU neural network and sentiment dictionary, to realize the tourism user evaluation semantic classification whole to tourism according to tourist's evaluation text whole to the travelling, comprising the following steps: 1) affective characteristics processing stage: carry out vectorization processing to the affective characteristics in tourism comment by constructing the compound dedicated sentiment dictionary of tourism；2) data preprocessing phase: to original comment text training term vector and context vector splicing is carried out, spliced vector is merged with the affective characteristics after vectorization, the input as two-way GRU neural network；3) the two-way GRU text semantic disaggregated model stage: the two-way GRU neural network of training simultaneously classifies to tourism evaluation emotion.Compared with prior art, the present invention has many advantages, such as that accuracy is high, considers the robustness of accuracy and the machine learning of sentiment dictionary.
Description
Technical field
The present invention relates to natural language processing and deep learning fields, more particularly, to a kind of combination At_GRU neural network
With the tourism evaluation sensibility classification method of sentiment dictionary.
Background technique
Travelling route evaluation is record tourist in the feedback for the specific itinerary that travel site formulates certain sight spot, is most
Directly satisfaction or suggestion opinion of the expression passenger to this itinerary, the vertical tie for connecting passenger and tour site.
It is evaluated by travelling route, passenger can elaborate the route and supply other trips to routing, lodging situation, transport arrangement etc.
Visitor uses for reference, and tourist corporation can also directly listen to opinions, and quick response improves, adjusts the details such as this itinerary, promotes clothes
Business, the satisfaction for reinforcing passenger.Therefore detailed analysis rapidly and accurately carried out to travelling route evaluation, handled and obtained precisely
Opinion rating and classification can be greatly improved itinerary optimal speed, shorten feedback gap, while reducing manual analysis expense
With and effectively promoted travel companies service quality.
Travelling route evaluation information is the important information for recording passenger's feedback, is mainly shown as the natural language of short text form
Epilegma is fallen.Simultaneously in recent years, be constantly in high speed development about natural language processing technique, especially using text semantic analyze as
Important research object is especially prominent, and text semantic analysis carries out structuring extraction, analysis and understanding to text, from semantic level
It is associated, thus accurate understanding text meaning.Semantic analysis includes conventional method and deep learning method, wherein depth
Study can extract more effective text features and higher accuracy rate compared to conventional method.At home and abroad have big
Amount scholar did correlative study, what has generation to propose that the combination of multiple features semanteme based on decision tree excavates, and summer name head et al. is used
ICTCLAS participle technique and word frequency statistics carry out the excavation of commodity evaluating characteristic, but fail to quote deep learning training, thus
More accurately feature cannot be extracted well, and Li Jie et al. carries out short text analysis using CNN model, but fails to make full use of
Hereafter semantic information causes not accurate enough problem of classifying.
Summary of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide a kind of combination At_GRU minds
Tourism evaluation sensibility classification method through network and sentiment dictionary.
The purpose of the present invention can be achieved through the following technical solutions:
A kind of tourism evaluation sensibility classification method of combination At_GRU neural network and sentiment dictionary, to according to tourist couple
The whole evaluation text of the travelling realizes the tourism user evaluation semantic classification whole to tourism, comprising the following steps:
1) affective characteristics processing stage: special to the emotion in tourism comment by constructing the compound dedicated sentiment dictionary of tourism
Sign carries out vectorization processing；
2) data preprocessing phase: to original comment text training term vector and context vector splicing is carried out, will be spliced
Vector afterwards is merged with the affective characteristics after vectorization, the input as two-way GRU neural network；
3) the two-way GRU text semantic disaggregated model stage: the two-way GRU neural network of training simultaneously carries out tourism evaluation emotion
Classification.
The step 1) specifically includes the following steps:
11) the compound dedicated sentiment dictionary of tourism is constructed: a variety of existing sentiment dictionaries of statistics, and merge therein identical
Polarity word forms compound sentiment dictionary, respectively to the tourism evaluation term vector after compound sentiment dictionary and participle, and obtains
The Euclidean distance for taking each word and all words of tourism evaluation in dictionary selects closely located as close similar emotion
Word, fusion form the dedicated sentiment dictionary of compound tourism；
12) semantic logic rule process: read tourism comment, using Chinese and English punctuation mark as subordinate sentence identify to comment on into
Row subordinate sentence, and obtain according to the part of speech of subordinate sentence the emotional value M (m of each subordinate sentence1,m2,m3,m4,m5), m1-m5The respectively subordinate sentence
In whether include negative word, degree adverb and emotion word, finally in conjunction with the emotion commented on as this of emotional value of each subordinate sentence
Polarity；
13) affective characteristics vectorization: the feeling polarities for the subordinate sentence handled well are carried out to increase dimension dyad.
In the step 11), existing sentiment dictionary includes Tsinghua University's dictionary, Hownet Hownet sentiment dictionary and platform
Gulf university simplified form of Chinese Character sentiment dictionary.
The step 2) specifically includes the following steps:
21) it segments, remove stop words and training term vector: original comment text information being defined as S, includes composition
Word collection W (w1,w2,...,wn), n is the word quantity of sentence S, after being segmented using ICTCLAS tool to original comment text simultaneously
Stop words is removed, using Word2vec tool to word collection W (w1,w2,...,wn) carry out term vector training, each word wiTable
It is shown as the term vector form of 50 dimensions；
22) it combines splicing context vector: defining i-th of word w in sentenceiAll remaining sentence information that the left side includes
For Cl(wi), the right residue sentence information is Cr(wi), be translated into after term vector generate combine spliced 50 dimension word to
Amount；
23) feature vector merges: the affective characteristics after spliced 50 dimension term vector and vectorization will be combined to count
Amount product processing fusion forms 55 final dimension term vectors, and as the input of GRU network.
In the step 22), expression formula is specifically defined are as follows:
cl(wi)=f (W(l)cl(wi-1)+W(sl)e(wi-1))
cr(wi)=f (W(r)cr(wi+1)+W(sr)e(wi+1))
Wherein, W(l)W(r)For by hidden layer, i.e. the context switch matrix that is next hidden layer, W(sl)For for combining
The matrix of current word and next word left side text relationship semanteme, W(sr)For for combining current word and a upper word
The matrix of the right text relationship semanteme, f are nonlinear activation function, cl(wi-1) it is (i-1)-th word wi-1The institute that the left side includes
There are remaining sentence information, e (wi-1) it is (i-1)-th word wi-1Term vector form, e (wi+1) it is i+1 word wi+1Word
Vector form, cr(wi+1) it is i+1 word wi+1All remaining sentence information that the right includes.
The step 3) specifically includes the following steps:
31) the two-way GRU network of training: constructing two-way GRU network, by 55 dimension term vector training sets respectively from sentence forward direction
It is loaded into At_GRU model with reversed, and carries out arameter optimization and complete training；
32) for having trained the two-way GRU network model completed, by new tourism user's evaluation carry out data prediction at
It for term vector, and is loaded into model and carries out emotional semantic classification, realize the natural language sentiment analysis to every user's evaluation, it is final to open up
Existing form be guide service, whether force consumption, traffic route, 5 dimensions of routing and lodging food and drink be satisfied with,
Generally, it is unsatisfied with the feeling polarities classification of three kinds of degree, and is indicated respectively with 1,0, -1, shows every travelling route passenger couple
The experience of 5 dimensions is fed back.
Compared with prior art, the invention has the following advantages that
One, by the exclusive dictionary of building tourism, the training corpus of great neural network is provided, it is logical compared to original
With sentiment dictionary, nicety of grading more can increase.
Two, the feelings that form reinforces context semantic relation and combination is generated based on semantic logic rule are spliced by term vector
Feel polar character and completely new feature vector is generated by specific fusion formula.Keep the accuracy of its existing sentiment dictionary also organic
The robustness of device study.
Three, by two-way GRU neural metwork training, compared to the unidirectional neural network of tradition, Direct/Reverse more can be parsed effectively
The semantic information of sentence.
Detailed description of the invention
Fig. 1 is semantic logic structure figure.
Fig. 2 is term vector context splicing construction figure.
Fig. 3 model framework structure flow chart.
Specific embodiment
The present invention proposes the tourism of a kind of pinpoint accuracy, the combination At_GRU neural network with study property and sentiment dictionary
Sensibility classification method is evaluated, as shown in figure 3, including three big steps:
One, affective characteristics processing stage
(1) tourism sentiment dictionary building: statistics Tsinghua University's dictionary, Hownet Hownet sentiment dictionary, Taiwan Univ. are simplified
A variety of sentiment dictionaries such as Chinese sentiment dictionary merge identical polar word, abundant opposed polarity word, improve and make compound feelings
Feel dictionary, later the tourism evaluation after the compound sentiment dictionary of term vectorization and participle respectively, and calculate in dictionary each word with
The Euclidean distance of all words of tourism evaluation, select it is closely located as close similar emotion word, and finally fusion make it is multiple
The dedicated sentiment dictionary of mould assembly tourism.
(2) semantic logic rule process: each tourism comment is read, with Chinese and English punctuation mark (.,?！！) make
Subordinate sentence is carried out to comment for subordinate sentence mark；Emotional value is obtained to each subordinate sentence of comment, process of calculation analysis is as shown in Figure 1, figure
In, Y indicates that emotional value be that 1, N indicates that emotional value is 0, to the emotional value of each subordinate sentence whether contain negative word, degree adverb,
Emotion word, exclamative sentence, confirmative question dimension are analyzed, and positive use " 1 " indicates, passive use " -1 " indicates, neutrality " 0 " table
Show, and forms the feeling polarities vector that 5 dimensional vectors are commented on as this.
(3) affective characteristics vectorization: the feeling polarities for the subordinate sentence handled well are carried out to increase dimension and term vector.
Two, data preprocessing phase
(1) participle, stop words removal and term vector training: urtext information is defined as S, the word collection comprising composition
W(w1,w2,...,wn), n represents the word quantity of sentence S.It is deactivated segmenting and removing to urtext using ICTCLAS tool
After word, using Word2vec tool to word collection W (w1,w2,...,wn) carry out term vector training, each word wiGenerate dimension
For 50 term vector form.
(2) it combines splicing context vector: defining i-th of word w in sentenceiAll remaining sentence information that the left side includes
For Cl(wi), the right residue sentence information is Cr(wi), be translated into after term vector generate combine spliced 50 dimension word to
Amount, specifically defines expression formula are as follows:
cl(wi)=f (W(l)cl(wi-1)+W(sl)e(wi-1))
cr(wi)=f (W(r)cr(wi+1)+W(sr)e(wi+1))
Wherein, W(l)W(r)For by hidden layer, i.e. the context switch matrix that is next hidden layer, W(sl)For for combining
The matrix of current word and next word left side text relationship semanteme, W(sr)For for combining current word and a upper word
The matrix of the right text relationship semanteme, f are nonlinear activation function, cl(wi-1) it is (i-1)-th word wi-1The institute that the left side includes
There are remaining sentence information, e (wi-1) it is (i-1)-th word wi-1Term vector form, e (wi+1) it is i+1 word wi+1Word
Vector form, cr(wi+1) it is i+1 word wi+1All remaining sentence information that the right includes；
Three, the two-way GRU text semantic disaggregated model stage
(1) two-way GRU network training: constructing two-way GRU network, by 55 dimension term vector training set Y { y1,y2,...,
ynBe loaded into At_GRU model from sentence forward direction and reversely respectively, form double-direction model training.
(2) system is loaded into the network model trained and completed, and is loaded into model for new tourism user's evaluation later, mould
Type carries out emotional semantic classification to it, realizes to every user's evaluation in route planning, whether forces consumption, traffic convenience, time peace
The dimensions such as row carry out the feeling polarities classification of satisfied, general, dissatisfied three kinds of degree, be tourist to the travelling route it is several compared with
There is clear and intuitive understanding in terms of for concern, better route decision is provided.
Embodiment
" one of domestic four overall gardens are proved to be for following tourism comment text！Either scale or setting, none is not
Embody the distinctive small bridge over the flowing stream in Jiangnan other, how beautiful pavilions, terraces and open halls is, and a winding path leads to quiet seclusion, and cobbled path is oblique！Lotus pool ... in water ".Implement
Following steps:
1, data prediction
1) data are segmented, the processing such as noise remove, is finally produced the following result:
" it deserve to be called domestic four overall gardens！Either scale setting, none embodies the peculiar small bridge over the flowing stream in Jiangnan, pavilions, terraces and open halls show
Beautiful, a winding path leads to quiet seclusion, and cobbled path is oblique！Lotus pool north lotus leaf is upper one times big in water, and lotus, which flickers, to blaze, more than ten days ... .. "
2) term vector training is carried out to each word, generates 50 dimension term vectors, as follows:
" gardens 0.15164 0.30177-0.16763 0.17684 0.31719 0.33973-0.43478-
0.31086-0.44999 -0.29486 0.16608 0.11963 -0.41328”
3) term vector for splicing each word and the associated word of its context, generates 50 completely new dimension term vectors
2, affective characteristics processing stage
1) be directed to already present several well-known sentiment dictionaries, as " Hownet homenet sentiment dictionary, Taiwan Univ. it is simplified in
Literary sentiment dictionary, Tsinghua University's sentiment dictionary etc. " is arranged and is integrated.
2) each word in term vector sentiment dictionary, and calculate and evaluate the European of term vector with what is generated in first step
Distance is selected with similarity in each sentiment dictionary close to 10% word, is formed and construct exclusive tourism sentiment dictionary.
3) it is based on exclusive tourism sentiment dictionary, the analysis and processing of feeling polarities are carried out to every tourism evaluation, analysis is
It is no if any mark " 1 ", not mark " 0 ", and ultimately generate 6 dimensional vectors there are also degree adverb, rhetorical question word etc., such as " 0,1,1.1,0,0,
1”。
4) term vector generated with first step carries out fusion treatment, final to generate 55 dimension term vector V1 " 0.31719
0.33973 -0.43478 -0.31086 -0.44999 -0.29486 0.16608 -0.41328....”。
3, the two-way GRU text semantic disaggregated model stage
1) GRU neural network is constructed, and is loaded into training set term vector V1
2) training network, and carry out artificial adjust and join, ultimately generate model M
3) it is loaded into test set, and finally generates classifying quality R (1,0,1, -1,0).
The present invention is spliced form by term vector and reinforces context semantic relation and combine to be generated based on semantic logic rule
Feeling polarities feature, pass through fusion and generate completely new feature vector.Making the accuracy of its existing sentiment dictionary also has engineering
The robustness of habit.The semantic classification method that the results show is proposed has higher accuracy.
Claims (6)
1. the tourism evaluation sensibility classification method of a kind of combination At_GRU neural network and sentiment dictionary, to according to tourist to this
The whole evaluation text of travelling realizes the tourism user evaluation semantic classification whole to tourism, which is characterized in that including following step
It is rapid:
1) affective characteristics processing stage: by construct the compound dedicated sentiment dictionary of tourism to the affective characteristics in tourism comment into
Row vectorization processing；
2) data preprocessing phase: to original comment text training term vector and carrying out context vector splicing, will be spliced
Vector is merged with the affective characteristics after vectorization, the input as two-way GRU neural network；
3) the two-way GRU text semantic disaggregated model stage: the two-way GRU neural network of training simultaneously divides tourism evaluation emotion
Class.
2. the tourism evaluation emotional semantic classification side of a kind of combination At_GRU neural network and sentiment dictionary according to claim 1
Method, which is characterized in that the step 1) specifically includes the following steps:
11) the compound dedicated sentiment dictionary of tourism is constructed: a variety of existing sentiment dictionaries of statistics, and merge identical polar therein
Word forms compound sentiment dictionary, respectively to the tourism evaluation term vector after compound sentiment dictionary and participle, and obtains word
The Euclidean distance of each word and all words of tourism evaluation in allusion quotation, selects closely located as close similar emotion word, melts
Conjunction forms the dedicated sentiment dictionary of compound tourism；
12) semantic logic rule process: tourism comment is read, is identified using Chinese and English punctuation mark as subordinate sentence and comment is divided
Sentence, and obtain according to the part of speech of subordinate sentence the emotional value of each subordinate sentence, in conjunction with the feelings commented on as this of emotional value of each subordinate sentence
Feel polarity；
13) affective characteristics vectorization: the feeling polarities for the subordinate sentence handled well are carried out to increase dimension dyad.
3. the tourism evaluation emotional semantic classification side of a kind of combination At_GRU neural network and sentiment dictionary according to claim 2
Method, which is characterized in that in the step 11), existing sentiment dictionary includes Tsinghua University's dictionary, Hownet Hownet emotion word
Allusion quotation and Taiwan Univ.'s simplified form of Chinese Character sentiment dictionary.
4. the tourism evaluation emotional semantic classification side of a kind of combination At_GRU neural network and sentiment dictionary according to claim 1
Method, which is characterized in that the step 2) specifically includes the following steps:
21) it segments, remove stop words and training term vector: original comment text information being defined as S, the word comprising composition
Collect W (w1,w2,...,wn), n is the word quantity of sentence S, after being segmented using ICTCLAS tool to original comment text and is removed
Stop words, using Word2vec tool to word collection W (w1,w2,...,wn) carry out term vector training, each word wiIt is expressed as
The term vector form of 50 dimensions；
22) it combines splicing context vector: defining i-th of word w in sentenceiAll remaining sentence information that the left side includes are Cl
(wi), the right residue sentence information is Cr(wi), it generates after being translated into term vector in conjunction with spliced 50 dimension term vector；
23) feature vector merges: the affective characteristics after spliced 50 dimension term vector and vectorization will be combined to carry out scalar product
Processing fusion forms 55 final dimension term vectors, and as the input of GRU network.
5. the tourism evaluation emotional semantic classification side of a kind of combination At_GRU neural network and sentiment dictionary according to claim 4
Method, which is characterized in that in the step 22), specifically define expression formula are as follows:
cl(wi)=f (W(l)cl(wi-1)+W(sl)e(wi-1))
cr(wi)=f (W(r)cr(wi+1)+W(sr)e(wi+1))
Wherein, W(l)W(r)For by hidden layer, i.e. the context switch matrix that is next hidden layer, W(sl)It is current for combining
The matrix of word and next word left side text relationship semanteme, W(sr)For for combining on the right of current word and a upper word
The matrix of text relationship semanteme, f are nonlinear activation function, cl(wi-1) it is (i-1)-th word wi-1The left side includes all surplus
Remaining sentence information, e (wi-1) it is (i-1)-th word wi-1Term vector form, e (wi+1) it is i+1 word wi+1Term vector
Form, cr(wi+1) it is i+1 word wi+1All remaining sentence information that the right includes.
6. the tourism evaluation emotional semantic classification side of a kind of combination At_GRU neural network and sentiment dictionary according to claim 1
Method, which is characterized in that the step 3) specifically includes the following steps:
31) the two-way GRU network of training: constructing two-way GRU network, by 55 dimension term vector training sets respectively from sentence forward direction and anti-
Into loading At_GRU model, and carries out arameter optimization and complete training；
32) for having trained the two-way GRU network model completed, new tourism user's evaluation, which is carried out data prediction, becomes word
Vector, and be loaded into model and carry out emotional semantic classification, it realizes to the natural language sentiment analysis of every user's evaluation, finally shows shape
Formula be guide service, whether force consumption, traffic route, 5 dimensions of routing and lodging food and drink carry out be satisfied with, one
As, the feeling polarities classification of dissatisfied three kinds of degree, and respectively with 1,0, -1 expression, show every travelling route passenger to this 5
The experience of a dimension is fed back.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810862476.4A CN109213861B (en) | 2018-08-01 | 2018-08-01 | Traveling evaluation emotion classification method combining At _ GRU neural network and emotion dictionary |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810862476.4A CN109213861B (en) | 2018-08-01 | 2018-08-01 | Traveling evaluation emotion classification method combining At _ GRU neural network and emotion dictionary |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109213861A true CN109213861A (en) | 2019-01-15 |
| CN109213861B CN109213861B (en) | 2022-03-29 |
Family
ID=64987780
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810862476.4A Active CN109213861B (en) | 2018-08-01 | 2018-08-01 | Traveling evaluation emotion classification method combining At _ GRU neural network and emotion dictionary |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109213861B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109858973A (en) * | 2019-02-18 | 2019-06-07 | 成都中科大旗软件有限公司 | A kind of analysis method of regional tourism industry development |
| CN109933664A (en) * | 2019-03-12 | 2019-06-25 | 中南大学 | A kind of fine granularity mood analysis improved method based on emotion word insertion |
| CN109948665A (en) * | 2019-02-28 | 2019-06-28 | 中国地质大学（武汉） | Physical activity genre classification methods and system based on long Memory Neural Networks in short-term |
| CN110019882A (en) * | 2019-03-18 | 2019-07-16 | 星潮闪耀移动网络科技（中国）有限公司 | A kind of advertising creative classification method and system |
| CN110059922A (en) * | 2019-03-11 | 2019-07-26 | 北京比速信息科技有限公司 | Satisfaction evaluation method on the line of data is commented on based on internet tourist |
| CN110083825A (en) * | 2019-03-21 | 2019-08-02 | 昆明理工大学 | A kind of Laotian sentiment analysis method based on GRU model |
| CN110196976A (en) * | 2019-05-10 | 2019-09-03 | 新华三大数据技术有限公司 | Sentiment orientation classification method, device and the server of text |
| CN110245299A (en) * | 2019-06-19 | 2019-09-17 | 中国人民解放军国防科技大学 | Sequence recommendation method and system based on dynamic interaction attention mechanism |
| CN110717090A (en) * | 2019-08-30 | 2020-01-21 | 昆山市量子昆慈量子科技有限责任公司 | Network public praise evaluation method and system for scenic spots and electronic equipment |
| CN111259651A (en) * | 2020-01-21 | 2020-06-09 | 北京工业大学 | User emotion analysis method based on multi-model fusion |
| WO2020199600A1 (en) * | 2019-04-03 | 2020-10-08 | 平安科技（深圳）有限公司 | Sentiment polarity analysis method and related device |
| CN112711664A (en) * | 2020-12-31 | 2021-04-27 | 山西三友和智慧信息技术股份有限公司 | Text emotion classification method based on TCN + LSTM |
| CN113033186A (en) * | 2021-05-31 | 2021-06-25 | 江苏联著实业股份有限公司 | Error correction early warning method and system based on event analysis |
| WO2021179455A1 (en) * | 2020-03-10 | 2021-09-16 | 深圳市华侨城创新研究院有限公司 | Scientometric method for measuring happiness indexes of tourist attractions |
| CN116533529A (en) * | 2023-05-12 | 2023-08-04 | 湖州东尼新能源有限公司 | Intelligent control method and system for ultrasonic welding PC (polycarbonate) sheet |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104572892A (en) * | 2014-12-24 | 2015-04-29 | 中国科学院自动化研究所 | Text classification method based on cyclic convolution network |
| CN107168945A (en) * | 2017-04-13 | 2017-09-15 | 广东工业大学 | A kind of bidirectional circulating neutral net fine granularity opinion mining method for merging multiple features |
| CN107609132A (en) * | 2017-09-18 | 2018-01-19 | 杭州电子科技大学 | One kind is based on Ontology storehouse Chinese text sentiment analysis method |
| CN108108433A (en) * | 2017-12-19 | 2018-06-01 | 杭州电子科技大学 | A kind of rule-based and the data network integration sentiment analysis method |
-
2018
- 2018-08-01 CN CN201810862476.4A patent/CN109213861B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104572892A (en) * | 2014-12-24 | 2015-04-29 | 中国科学院自动化研究所 | Text classification method based on cyclic convolution network |
| CN107168945A (en) * | 2017-04-13 | 2017-09-15 | 广东工业大学 | A kind of bidirectional circulating neutral net fine granularity opinion mining method for merging multiple features |
| CN107609132A (en) * | 2017-09-18 | 2018-01-19 | 杭州电子科技大学 | One kind is based on Ontology storehouse Chinese text sentiment analysis method |
| CN108108433A (en) * | 2017-12-19 | 2018-06-01 | 杭州电子科技大学 | A kind of rule-based and the data network integration sentiment analysis method |
Non-Patent Citations (1)
| Title |
|---|
| 李科: "基于多元特征融合和 LSTM神经网络的中文评论情感分析", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109858973A (en) * | 2019-02-18 | 2019-06-07 | 成都中科大旗软件有限公司 | A kind of analysis method of regional tourism industry development |
| CN109948665A (en) * | 2019-02-28 | 2019-06-28 | 中国地质大学（武汉） | Physical activity genre classification methods and system based on long Memory Neural Networks in short-term |
| CN110059922A (en) * | 2019-03-11 | 2019-07-26 | 北京比速信息科技有限公司 | Satisfaction evaluation method on the line of data is commented on based on internet tourist |
| CN109933664A (en) * | 2019-03-12 | 2019-06-25 | 中南大学 | A kind of fine granularity mood analysis improved method based on emotion word insertion |
| CN110019882A (en) * | 2019-03-18 | 2019-07-16 | 星潮闪耀移动网络科技（中国）有限公司 | A kind of advertising creative classification method and system |
| CN110083825A (en) * | 2019-03-21 | 2019-08-02 | 昆明理工大学 | A kind of Laotian sentiment analysis method based on GRU model |
| WO2020199600A1 (en) * | 2019-04-03 | 2020-10-08 | 平安科技（深圳）有限公司 | Sentiment polarity analysis method and related device |
| CN110196976A (en) * | 2019-05-10 | 2019-09-03 | 新华三大数据技术有限公司 | Sentiment orientation classification method, device and the server of text |
| CN110245299A (en) * | 2019-06-19 | 2019-09-17 | 中国人民解放军国防科技大学 | Sequence recommendation method and system based on dynamic interaction attention mechanism |
| CN110245299B (en) * | 2019-06-19 | 2022-02-08 | 中国人民解放军国防科技大学 | Sequence recommendation method and system based on dynamic interaction attention mechanism |
| CN110717090A (en) * | 2019-08-30 | 2020-01-21 | 昆山市量子昆慈量子科技有限责任公司 | Network public praise evaluation method and system for scenic spots and electronic equipment |
| CN111259651A (en) * | 2020-01-21 | 2020-06-09 | 北京工业大学 | User emotion analysis method based on multi-model fusion |
| WO2021179455A1 (en) * | 2020-03-10 | 2021-09-16 | 深圳市华侨城创新研究院有限公司 | Scientometric method for measuring happiness indexes of tourist attractions |
| CN112711664A (en) * | 2020-12-31 | 2021-04-27 | 山西三友和智慧信息技术股份有限公司 | Text emotion classification method based on TCN + LSTM |
| CN112711664B (en) * | 2020-12-31 | 2022-09-20 | 山西三友和智慧信息技术股份有限公司 | Text emotion classification method based on TCN + LSTM |
| CN113033186A (en) * | 2021-05-31 | 2021-06-25 | 江苏联著实业股份有限公司 | Error correction early warning method and system based on event analysis |
| CN113033186B (en) * | 2021-05-31 | 2021-08-17 | 江苏联著实业股份有限公司 | Error correction early warning method and system based on event analysis |
| CN116533529A (en) * | 2023-05-12 | 2023-08-04 | 湖州东尼新能源有限公司 | Intelligent control method and system for ultrasonic welding PC (polycarbonate) sheet |
| CN116533529B (en) * | 2023-05-12 | 2023-09-29 | 湖州东尼新能源有限公司 | Intelligent control method and system for ultrasonic welding PC (polycarbonate) sheet |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109213861B (en) | 2022-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109213861A (en) | In conjunction with the tourism evaluation sensibility classification method of At_GRU neural network and sentiment dictionary | |
| Liu et al. | Learning to assemble neural module tree networks for visual grounding | |
| CN110633409B (en) | Automobile news event extraction method integrating rules and deep learning | |
| CN103631961B (en) | Method for identifying relationship between sentiment words and evaluation objects | |
| CN107168945A (en) | A kind of bidirectional circulating neutral net fine granularity opinion mining method for merging multiple features | |
| CN111581474B (en) | Evaluation object extraction method of case-related microblog comments based on multi-head attention system | |
| CN109933664A (en) | A kind of fine granularity mood analysis improved method based on emotion word insertion | |
| CN107092596A (en) | Text emotion analysis method based on attention CNNs and CCR | |
| CN105930503A (en) | Combination feature vector and deep learning based sentiment classification method and device | |
| CN106407235B (en) | A kind of semantic dictionary construction method based on comment data | |
| CN104794212A (en) | Context sentiment classification method and system based on user comment text | |
| CN108563638B (en) | Microblog emotion analysis method based on topic identification and integrated learning | |
| CN108536870A (en) | A kind of text sentiment classification method of fusion affective characteristics and semantic feature | |
| CN110609983B (en) | Structured decomposition method for policy file | |
| CN109948158A (en) | Emotional orientation analytical method based on environment member insertion and deep learning | |
| CN106126751A (en) | A kind of sorting technique with time availability and device | |
| CN109670039A (en) | Sentiment analysis method is commented on based on the semi-supervised electric business of tripartite graph and clustering | |
| CN108419123A (en) | A kind of virtual sliced sheet method of instructional video | |
| CN109949637B (en) | Automatic answering method and device for objective questions | |
| CN108932322A (en) | A kind of geographical semantics method for digging based on text big data | |
| CN109492105A (en) | A kind of text sentiment classification method based on multiple features integrated study | |
| CN110119443A (en) | A kind of sentiment analysis method towards recommendation service | |
| CN110263332A (en) | A kind of natural language Relation extraction method neural network based | |
| CN105975639A (en) | Search result ordering method and device | |
| CN110147552A (en) | Educational resource quality evaluation method for digging and system based on natural language processing |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |