CN109299268A - A kind of text emotion analysis method based on dual channel model - Google Patents
A kind of text emotion analysis method based on dual channel model Download PDFInfo
- Publication number
- CN109299268A CN109299268A CN201811239484.XA CN201811239484A CN109299268A CN 109299268 A CN109299268 A CN 109299268A CN 201811239484 A CN201811239484 A CN 201811239484A CN 109299268 A CN109299268 A CN 109299268A
- Authority
- CN
- China
- Prior art keywords
- text
- term vector
- feature
- cnn
- layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Abstract
The present invention is single for the problem that current single channel Artificial Neural Network Structures, is unable to fully extract text information, proposes a kind of text emotion analysis method based on dual channel model.Comprising steps of first using Word2Vec training term vector, by text representation at the form of term vector matrix；Then it as input data, is respectively fed to carry out feature extraction in convolutional neural networks (CNN) and long short-term memory (LSTM) network；And attention model is introduced behind, extract text important feature information；Finally two extracted text features in channel are merged, carry out emotional semantic classification using classification layer.The mentioned method of the present invention has feasibility and superiority, and performance is substantially better than other single channel neural network models.
Description
Technical field
The present invention proposes a kind of text emotion analysis method based on dual channel model, is related to text emotion analysis field.
Background technique
In recent years, with the fast development of Internet industry, there are numerous New Medias, these new medias constantly rush
Hit and change people's lives mode.The emergence of various electric business platforms, so that the shopping at network stayed indoors becomes simple
Prevalence, and unique feedback that this shopping way obtains is exactly the comment of consumption experience left by consumer, these are really commented
By being to determine unique foundation for whether consuming of potential consumer.Therefore the text that these are largely had with emotion carries out emotion
Analysis, is the work all beneficial to electric business platform or Consumer groups.
The main task of text emotion analysis is exactly to analyze the text information with emotion, extracts feature,
And make polarity judgement.There are mainly two types of the methods of text emotion analysis at present, method based on sentiment dictionary and is based on machine
The method of study.Wherein calculated based on the method for emotion dictionary by carrying out the combination of certain way to the emotion word in text,
Finally obtain the feeling polarities of text.But with the development of society, this method face the increasingly text of diversification when, effect
With regard to not ideal enough.In addition, carrying out text emotion analysis using machine learning method, the data characteristics of a large amount of engineers is needed,
With the increase of text data set to be processed, the deep layer for the learning text that conventional machines learning method can not be faster and better
Information characteristics.
It the use of conventional method when carrying out text information feature extraction, is all being that indiscriminate feature is carried out to text
It extracts, it would be appreciated that each word is different the contribution of entire text feeling polarities in text, common mind
Through network, such as convolutional neural networks (Convolutional Neural Network, CNN) and long memory network (Long in short-term
Short-Term Memory, LSTM), the position of important information in sentence can not be differentiated when extracting feature, so as to cause text
Classification results are influenced very big by irrelevant information.Existing sentiment analysis method is substantially by constructing single pass nerve net
Network model carries out the extraction of text feature, and one channel model, with the increase of the network number of plies, performance will be affected, cannot
Enough features for adequately extracting text.
The present invention proposes a kind of text emotion analysis method based on dual channel model, constructs a kind of binary channels attention
Model (Dual-Channel Attention Model, DCAM), first using Word2Vec training term vector, by text representation
At the form of term vector matrix；Then as input data, convolutional neural networks (CNN) and long short-term memory are respectively fed to
(LSTM) feature extraction is carried out in network；And attention model is introduced behind, extract text important feature information；Finally by two
A extracted text feature in channel merges, and carries out emotional semantic classification using classification layer.This method is by the advantage of CNN and LSTM
It integrates, that is, is extracted the local feature of text, it is also considered that the sequence information of text, and can more fully mention
Text feature information is taken, enhances important information to the influence power of text, reduces interference of the inessential information to text classification, thus
Promote classifying quality.
Summary of the invention
In view of this, it is a primary object of the present invention to integrate the advantage of CNN and LSTM, and respectively at two
Attention mechanism is introduced behind channel, is enhanced the resolution capability to important information, is reduced interference of the inessential information to text classification,
To promote classification accuracy rate.
In order to achieve the above object, technical solution proposed by the present invention are as follows:
A kind of text emotion analysis method based on dual channel model, described method includes following steps:
Step 1 pre-processes data set, using Word2Vec training term vector, by text representation at term vector matrix
Form；
Step 2, using trained term vector matrix as the input feature vector of CNN, and carry out convolution operation, learning text part is special
Sign, and attention layer is accessed behind；
Step 3, using trained term vector matrix as the input of LSTM network, the sequence information of learning text, and use note
Power mechanism of anticipating learns important text information；
Step 4, the text feature for extracting two channels merge, and input Softmax classification layer is classified, by with text
Original label comparing calculation, obtains text classification accuracy.
In conclusion the present invention combines the sequence information learning ability of the local shape factor ability LSTM of CNN, and
Identify that important feature, enhancing important information reduce inessential information to text point to the influence power of text using attention mechanism
The interference of class, to promote classification accuracy rate.
Detailed description of the invention:
Fig. 1 is a kind of overall procedure schematic diagram of the text emotion analysis method based on dual channel model of the present invention；
Fig. 2 is using Word2Vec training term vector, by text representation at the flow diagram of term vector matrix form；
Fig. 3 is that feature flow diagram is extracted in the channel CNN；
Fig. 4 is that feature flow diagram is extracted in the channel LSTM；
Fig. 5 is to be classified using Softmax classification layer, calculates the flow diagram of text classification accuracy；
Fig. 6 is to use the obtained accuracy rate result of Chinese data collection.
Specific embodiment:
Below in conjunction with attached drawing of the invention, technical solution of the present invention is clearly and completely described, it is clear that lift real
Example is for illustrating, and non-limiting embodiments of the present invention, the present invention can also pass through other different specific embodiment parties
Formula is implemented.Every other embodiment obtained by those of ordinary skill in the art without making creative efforts, all
Belong to the scope of protection of the invention.
Fig. 1 is a kind of overall procedure signal of text emotion analysis method based on dual channel model of the present invention
Figure, as shown in Figure 1, the text emotion analysis method of the present invention based on dual channel model, includes the following steps:
Step 1 pre-processes data set, using Word2Vec training term vector, by text representation at term vector matrix
Form
Step 2, using trained term vector matrix as the input feature vector of CNN, and carry out convolution operation, learning text part is special
Sign, and attention layer is accessed behind；
Step 3, using trained term vector matrix as the input of LSTM network, the sequence information of learning text, and use note
Power mechanism of anticipating learns important text information；
Step 4, the text feature for extracting two channels merge, and input Softmax classification layer is classified, by with text
Original label comparing calculation, obtains text classification accuracy.
Fig. 2 is using Word2Vec training term vector, by text representation at the flow diagram of term vector matrix form, such as
Shown in Fig. 2, in step 1, data set is pre-processed, using Word2Vec training term vector, by text representation at term vector
The form of matrix, includes the following steps:
Step 11 pre-processes text, carries out word segmentation processing to corpus using stammerer participle tool, and filter out deactivated
Word；
Pretreated corpus is trained for fixed dimension term vector using Word2Vec by step 12, and format is as follows:, wherein d represents term vector dimension, value 128；
Term vector is spliced the term vector matrix for becoming fixed size by step 13 in sequence.
Fig. 3 is the channel CNN extraction feature flow diagram, as shown in figure 3, in step 2, by trained term vector matrix
As the input feature vector of CNN, and convolution operation is carried out, learning text local feature, and attention layer is accessed behind, including such as
Lower step:
Step 21 executes convolution operation by way of using linear filter on input matrix, can be expressed as follows:
In above formulaFor bias term,To be highly=3,4,5, width is=128 convolution kernel,Swash to be non-linear
Function living is expressed as following formula using ReLu as activation primitive:
After the completion of step 22, convolution operation, a characteristic pattern C is obtained,
In above formula, branch indicates the feature generated to convolution operationSpliced;
Step 23, the feature for being obtained CNN using tanh functionNonlinear transformation is, it is expressed as；
Step 24, initialization attention matrix, by its withPoint multiplication operation is carried out, it is carried out using Softmax function
Normalization operation obtains CNN layers of output weight coefficient,
Step 25, by by weight coefficientWith characteristic patternIt is multiplied, it is special to obtain CNN extracted text after attention layer
Sign, is expressed as vector。
Fig. 4 is the channel LSTM extraction feature flow diagram, as shown in figure 4, in step 3, by trained term vector square
Input of the battle array as LSTM network, the sequence information of learning text, and learn important text information using attention mechanism, including
Following steps:
The term vector matrix of converted fixed size is divided by step 31A time step, i.e., each time step handle one
The term vector of word is sent to LSTM network and is trained, and finally obtains the current hidden layer state of LSTM network
Step 32, by the hiding layer state in LSTM networkNonlinear transformation is:
Step 33, initialization attention matrix, by its withPoint multiplication operation is carried out, it is returned using Softmax function
One changes operation, obtains LSTM layers of output weight coefficient:
Step 34, by the way that weight coefficient is multiplied with hidden layer state, obtain the feature for the LSTM layer extraction for introducing attention mechanism
Vector, it is expressed as。
Fig. 5 is to be classified using Softmax classification layer, the flow diagram of text classification accuracy is calculated, such as Fig. 5 institute
Show, in step 4, the text feature that two channels are extracted merges, and input Softmax classification layer is classified, by with text
This original label comparing calculation, obtains text classification accuracy, includes the following steps:
Step 41, the feature for extracting the CNN layer for introducing attention mechanismIt is extracted with the LSTM layer for introducing attention mechanism
Feature vectorFusion
Step 42, using Dropout technology, when calculating, random selection inactivates a part of neuron；
Step 43 is calculated each text using Softmax function and adheres to different classes of probability size p separately；
Step 44 carries out text categories judgement, is judged to biggish classification belonging to probability value p, calculates classification accuracy rate.
Embodiment:
This example uses the true Chinese comment acquired from internet, uses the text emotion analysis method based on dual channel model
Text emotion is analyzed, it is shown that specific step is as follows:
1. a pair data set pre-processes, segmented using stammerer and carry out word segmentation processing, remove stop words, is 60 by text fixed length；
2. positive emotion text assigns label 1, Negative Affect text assigns label 0, partition testing collection and training set；
3. using Word2Vec tool training term vector, term vector is spliced by dimension set 128 according to text word sequence
60128 term vector matrix；
4. using term vector matrix as the input feature vector of CNN and LSTM network, wherein the convolution kernel of CNN is dimensioned to
3,4,5, the hidden neuron number of number 128, LSTM network is set as 128；
5. accessing attention layer after CNN and LSTM network respectively extracts important feature information, attention layer size and the channel
Extracted feature sizes are consistent；
6. the obtained feature in two channels is accessed full articulamentum, finally enters classification layer and classify, obtain such as Fig. 6 institute
The accuracy shown is as a result, horizontal axis represents experiment the number of iterations, and the longitudinal axis represents accuracy, wherein being labeled as the representative of equilateral triangle
CNN-Attention model, the model use single channel CNN network and introduce attention mechanism behind；Labeled as inverted triangle
The representative LSTM-Attention model of shape, the model use single channel LSTM network and introduce attention mechanism behind；Mark
It is denoted as the representative DC-CNN-LSTM model of five-pointed star, which carries out binary channels feature extraction using CNN and LSTM network；Mark
Being denoted as hexagon is model constructed by method of the invention, and accuracy is up to 92.7% on test set.
Claims (5)
1. a kind of text emotion analysis method based on dual channel model, which is characterized in that the text emotion analysis method packet
Include following steps:
Step 1 pre-processes data set, using Word2Vec training term vector, by text representation at term vector matrix
Form；
Step 2, using trained term vector matrix as the input feature vector of CNN, and carry out convolution operation, learning text part is special
Sign, and attention layer is accessed behind；
Step 3, using trained term vector matrix as the input of LSTM network, the sequence information of learning text, and use note
Power mechanism of anticipating learns important text information；
Step 4, the text feature for extracting two channels merge, and input Softmax classification layer is classified, by with text
Original label comparing calculation, obtains text classification accuracy.
2. a kind of text emotion analysis method based on dual channel model according to claim 1, which is characterized in that step
In 1, data set is pre-processed, and using Word2Vec training term vector, by text representation at the form of term vector matrix；
Include the following steps:
Step 11 pre-processes text, carries out word segmentation processing to corpus using stammerer participle tool, and filter out deactivated
Word；
Pretreated corpus is trained for fixed dimension term vector using Word2Vec by step 12, and format is as follows:, wherein d represents term vector dimension, value 128；
Term vector is spliced the term vector matrix for becoming fixed size by step 13 in sequence.
3. a kind of text emotion analysis method based on dual channel model according to claim 1, which is characterized in that in step
In rapid 2, using trained term vector matrix as the input feature vector of CNN, and convolution operation is carried out, learning text local feature,
And attention layer is accessed behind；Include the following steps:
Step 21 executes convolution operation by way of using linear filter on input matrix, can be expressed as follows:
In above formulaFor bias term,To be highly=3,4,5, width is=128 convolution kernel,For nonlinear activation
Function is expressed as following formula using ReLu function as activation primitive:
After the completion of step 22, convolution operation, a characteristic pattern C is obtained:
In above formula, branch indicates the feature generated to convolution operationSpliced;
Step 23, the feature for being obtained CNN using tanh functionNonlinear transformation is, it is expressed as；
Step 24, initialization attention matrix, by its withPoint multiplication operation is carried out, it is returned using Softmax function
One changes operation, obtains CNN layers of output weight coefficient,
Step 25, by by weight coefficientWith characteristic patternIt is multiplied, it is special to obtain CNN extracted text after attention layer
Sign, is expressed as vector。
4. a kind of text emotion analysis method based on dual channel model according to claim 1, which is characterized in that in step
In rapid 3, using trained term vector matrix as the input of LSTM network, the sequence information of learning text, and attention is used
Mechanism learns important text information, includes the following steps:
The term vector matrix of converted fixed size is divided by step 31=60 time steps, i.e., each time step processing one
The term vector of a word is sent to LSTM network and is trained, and finally obtains the current hidden layer state of LSTM network；
Step 32, using tanh function by the hiding layer state in LSTM networkNonlinear transformation is；
Step 33, initialization attention matrix, by its withPoint multiplication operation is carried out, it is returned using Softmax function
One changes operation, obtains LSTM layers of output weight coefficient；
Step 34, by the way that weight coefficient is multiplied with hidden layer state, obtain the feature for the LSTM layer extraction for introducing attention mechanism
Vector, it is expressed as。
5. a kind of text emotion analysis method based on dual channel model according to claim 1, which is characterized in that step
In 4, the text feature that two channels are extracted merges, and input Softmax classification layer is classified, by with the original mark of text
Comparing calculation is signed, obtains text classification accuracy；Include the following steps:
Step 41, the feature for extracting the CNN layer for introducing attention mechanismThe spy extracted with the LSTM layer for introducing attention mechanism
Levy vectorFusion, accesses full articulamentum；
Step 42, using Dropout technology, when calculating, random selection inactivates a part of neuron；
Step 43 adheres to different classes of probability size p separately to calculate each text using being Softmax function；
Step 44 carries out text categories judgement, is judged to biggish classification belonging to probability value p, calculates classification accuracy rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811239484.XA CN109299268A (en) | 2018-10-24 | 2018-10-24 | A kind of text emotion analysis method based on dual channel model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811239484.XA CN109299268A (en) | 2018-10-24 | 2018-10-24 | A kind of text emotion analysis method based on dual channel model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109299268A true CN109299268A (en) | 2019-02-01 |
Family
ID=65157711
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811239484.XA Pending CN109299268A (en) | 2018-10-24 | 2018-10-24 | A kind of text emotion analysis method based on dual channel model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109299268A (en) |
Cited By (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109902177A (en) * | 2019-02-28 | 2019-06-18 | 上海理工大学 | Text emotion analysis method based on binary channels convolution Memory Neural Networks |
| CN110008338A (en) * | 2019-03-04 | 2019-07-12 | 华南理工大学 | A kind of electric business evaluation sentiment analysis method of fusion GAN and transfer learning |
| CN110032645A (en) * | 2019-04-17 | 2019-07-19 | 携程旅游信息技术（上海）有限公司 | Text emotion recognition methods, system, equipment and medium |
| CN110110330A (en) * | 2019-04-30 | 2019-08-09 | 腾讯科技（深圳）有限公司 | Text based keyword extracting method and computer equipment |
| CN110222149A (en) * | 2019-05-17 | 2019-09-10 | 华中科技大学 | A kind of Time Series Forecasting Methods based on news public sentiment |
| CN110232123A (en) * | 2019-05-28 | 2019-09-13 | 第四范式（北京）技术有限公司 | The sentiment analysis method and device thereof of text calculate equipment and readable medium |
| CN110263152A (en) * | 2019-05-07 | 2019-09-20 | 平安科技（深圳）有限公司 | File classification method, system and computer equipment neural network based |
| CN110263328A (en) * | 2019-05-22 | 2019-09-20 | 平安国际智慧城市科技股份有限公司 | A kind of disciplinary capability type mask method, device, storage medium and terminal device |
| CN110275954A (en) * | 2019-06-21 | 2019-09-24 | 深圳大学 | Chinese Text Categorization based on time convolutional network |
| CN110321565A (en) * | 2019-07-09 | 2019-10-11 | 广东工业大学 | Real-time text sentiment analysis method, device and equipment based on deep learning |
| CN110413785A (en) * | 2019-07-25 | 2019-11-05 | 淮阴工学院 | A kind of Automatic document classification method based on BERT and Fusion Features |
| CN110459241A (en) * | 2019-08-30 | 2019-11-15 | 厦门亿联网络技术股份有限公司 | A kind of extracting method and system for phonetic feature |
| CN110490813A (en) * | 2019-07-05 | 2019-11-22 | 特斯联（北京）科技有限公司 | Characteristic pattern Enhancement Method, device, equipment and the medium of convolutional neural networks |
| CN110502757A (en) * | 2019-08-29 | 2019-11-26 | 西安邮电大学 | A kind of natural language sentiment analysis method |
| CN110855474A (en) * | 2019-10-21 | 2020-02-28 | 广州杰赛科技股份有限公司 | Network feature extraction method, device, equipment and storage medium of KQI data |
| CN110874410A (en) * | 2019-11-01 | 2020-03-10 | 河南理工大学 | Text classification method based on long-time and short-time memory network and convolutional neural network |
| CN111061873A (en) * | 2019-11-28 | 2020-04-24 | 北京工业大学 | Multi-channel text classification method based on Attention mechanism |
| CN111078822A (en) * | 2019-11-29 | 2020-04-28 | 北京百卓网络技术有限公司 | Reader information extraction method and system based on Chinese novel text |
| CN111078833A (en) * | 2019-12-03 | 2020-04-28 | 哈尔滨工程大学 | Text classification method based on neural network |
| CN111147396A (en) * | 2019-12-26 | 2020-05-12 | 哈尔滨工程大学 | Encrypted flow classification method based on sequence characteristics |
| CN111324734A (en) * | 2020-02-17 | 2020-06-23 | 昆明理工大学 | Case microblog comment emotion classification method integrating emotion knowledge |
| CN111339768A (en) * | 2020-02-27 | 2020-06-26 | 携程旅游网络技术（上海）有限公司 | Sensitive text detection method, system, electronic device and medium |
| CN111401473A (en) * | 2020-04-09 | 2020-07-10 | 中国人民解放军国防科技大学 | Infrared target classification method based on attention mechanism convolutional neural network |
| CN111476408A (en) * | 2020-03-26 | 2020-07-31 | 国网江苏省电力有限公司信息通信分公司 | Power communication equipment state prediction method and system |
| CN111522956A (en) * | 2020-05-08 | 2020-08-11 | 河南理工大学 | Text emotion classification method based on double channels and hierarchical attention network |
| CN111651593A (en) * | 2020-05-08 | 2020-09-11 | 河南理工大学 | Text emotion analysis method based on word vector and word vector mixed model |
| CN111651603A (en) * | 2020-06-04 | 2020-09-11 | 上海电力大学 | Power industry single-level text classification method and system based on LAV parameter fine adjustment |
| CN111696624A (en) * | 2020-06-08 | 2020-09-22 | 天津大学 | DNA binding protein identification and function annotation deep learning method based on self-attention mechanism |
| CN111881262A (en) * | 2020-08-06 | 2020-11-03 | 重庆邮电大学 | Text emotion analysis method based on multi-channel neural network |
| CN112818124A (en) * | 2021-02-21 | 2021-05-18 | 昆明理工大学 | Entity relationship extraction method based on attention neural network |
| CN113191138A (en) * | 2021-05-14 | 2021-07-30 | 长江大学 | Automatic text emotion analysis method based on AM-CNN algorithm |
| CN113326374A (en) * | 2021-05-25 | 2021-08-31 | 成都信息工程大学 | Short text emotion classification method and system based on feature enhancement |
| CN114022909A (en) * | 2022-01-07 | 2022-02-08 | 首都师范大学 | Emotion recognition method and system based on sensor data |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609009A (en) * | 2017-07-26 | 2018-01-19 | 北京大学深圳研究院 | Text emotion analysis method, device, storage medium and computer equipment |
| CN108415972A (en) * | 2018-02-08 | 2018-08-17 | 合肥工业大学 | text emotion processing method |
| CN108427670A (en) * | 2018-04-08 | 2018-08-21 | 重庆邮电大学 | A kind of sentiment analysis method based on context word vector sum deep learning |
-
2018
- 2018-10-24 CN CN201811239484.XA patent/CN109299268A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609009A (en) * | 2017-07-26 | 2018-01-19 | 北京大学深圳研究院 | Text emotion analysis method, device, storage medium and computer equipment |
| CN108415972A (en) * | 2018-02-08 | 2018-08-17 | 合肥工业大学 | text emotion processing method |
| CN108427670A (en) * | 2018-04-08 | 2018-08-21 | 重庆邮电大学 | A kind of sentiment analysis method based on context word vector sum deep learning |
Cited By (51)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109902177A (en) * | 2019-02-28 | 2019-06-18 | 上海理工大学 | Text emotion analysis method based on binary channels convolution Memory Neural Networks |
| CN109902177B (en) * | 2019-02-28 | 2022-11-29 | 上海理工大学 | Text emotion analysis method based on dual-channel convolutional memory neural network |
| CN110008338A (en) * | 2019-03-04 | 2019-07-12 | 华南理工大学 | A kind of electric business evaluation sentiment analysis method of fusion GAN and transfer learning |
| CN110032645A (en) * | 2019-04-17 | 2019-07-19 | 携程旅游信息技术（上海）有限公司 | Text emotion recognition methods, system, equipment and medium |
| CN110032645B (en) * | 2019-04-17 | 2021-02-09 | 携程旅游信息技术（上海）有限公司 | Text emotion recognition method, system, device and medium |
| CN110110330A (en) * | 2019-04-30 | 2019-08-09 | 腾讯科技（深圳）有限公司 | Text based keyword extracting method and computer equipment |
| CN110110330B (en) * | 2019-04-30 | 2023-08-11 | 腾讯科技（深圳）有限公司 | Keyword extraction method based on text and computer equipment |
| CN110263152B (en) * | 2019-05-07 | 2024-04-09 | 平安科技（深圳）有限公司 | Text classification method, system and computer equipment based on neural network |
| CN110263152A (en) * | 2019-05-07 | 2019-09-20 | 平安科技（深圳）有限公司 | File classification method, system and computer equipment neural network based |
| CN110222149A (en) * | 2019-05-17 | 2019-09-10 | 华中科技大学 | A kind of Time Series Forecasting Methods based on news public sentiment |
| CN110263328A (en) * | 2019-05-22 | 2019-09-20 | 平安国际智慧城市科技股份有限公司 | A kind of disciplinary capability type mask method, device, storage medium and terminal device |
| CN110263328B (en) * | 2019-05-22 | 2024-01-23 | 平安国际智慧城市科技股份有限公司 | Discipline capability type labeling method and device, storage medium and terminal equipment |
| CN110232123A (en) * | 2019-05-28 | 2019-09-13 | 第四范式（北京）技术有限公司 | The sentiment analysis method and device thereof of text calculate equipment and readable medium |
| CN110275954A (en) * | 2019-06-21 | 2019-09-24 | 深圳大学 | Chinese Text Categorization based on time convolutional network |
| CN110490813A (en) * | 2019-07-05 | 2019-11-22 | 特斯联（北京）科技有限公司 | Characteristic pattern Enhancement Method, device, equipment and the medium of convolutional neural networks |
| CN110321565B (en) * | 2019-07-09 | 2024-02-23 | 广东工业大学 | Real-time text emotion analysis method, device and equipment based on deep learning |
| CN110321565A (en) * | 2019-07-09 | 2019-10-11 | 广东工业大学 | Real-time text sentiment analysis method, device and equipment based on deep learning |
| CN110413785A (en) * | 2019-07-25 | 2019-11-05 | 淮阴工学院 | A kind of Automatic document classification method based on BERT and Fusion Features |
| CN110413785B (en) * | 2019-07-25 | 2021-10-19 | 淮阴工学院 | Text automatic classification method based on BERT and feature fusion |
| CN110502757B (en) * | 2019-08-29 | 2023-01-10 | 西安邮电大学 | Natural language emotion analysis method |
| CN110502757A (en) * | 2019-08-29 | 2019-11-26 | 西安邮电大学 | A kind of natural language sentiment analysis method |
| CN110459241A (en) * | 2019-08-30 | 2019-11-15 | 厦门亿联网络技术股份有限公司 | A kind of extracting method and system for phonetic feature |
| CN110459241B (en) * | 2019-08-30 | 2022-03-04 | 厦门亿联网络技术股份有限公司 | Method and system for extracting voice features |
| CN110855474A (en) * | 2019-10-21 | 2020-02-28 | 广州杰赛科技股份有限公司 | Network feature extraction method, device, equipment and storage medium of KQI data |
| CN110855474B (en) * | 2019-10-21 | 2022-06-17 | 广州杰赛科技股份有限公司 | Network feature extraction method, device, equipment and storage medium of KQI data |
| CN110874410A (en) * | 2019-11-01 | 2020-03-10 | 河南理工大学 | Text classification method based on long-time and short-time memory network and convolutional neural network |
| CN111061873A (en) * | 2019-11-28 | 2020-04-24 | 北京工业大学 | Multi-channel text classification method based on Attention mechanism |
| CN111078822A (en) * | 2019-11-29 | 2020-04-28 | 北京百卓网络技术有限公司 | Reader information extraction method and system based on Chinese novel text |
| CN111078833B (en) * | 2019-12-03 | 2022-05-20 | 哈尔滨工程大学 | Text classification method based on neural network |
| CN111078833A (en) * | 2019-12-03 | 2020-04-28 | 哈尔滨工程大学 | Text classification method based on neural network |
| CN111147396B (en) * | 2019-12-26 | 2023-03-21 | 哈尔滨工程大学 | Encrypted flow classification method based on sequence characteristics |
| CN111147396A (en) * | 2019-12-26 | 2020-05-12 | 哈尔滨工程大学 | Encrypted flow classification method based on sequence characteristics |
| CN111324734A (en) * | 2020-02-17 | 2020-06-23 | 昆明理工大学 | Case microblog comment emotion classification method integrating emotion knowledge |
| CN111339768B (en) * | 2020-02-27 | 2024-03-05 | 携程旅游网络技术（上海）有限公司 | Sensitive text detection method, system, electronic equipment and medium |
| CN111339768A (en) * | 2020-02-27 | 2020-06-26 | 携程旅游网络技术（上海）有限公司 | Sensitive text detection method, system, electronic device and medium |
| CN111476408A (en) * | 2020-03-26 | 2020-07-31 | 国网江苏省电力有限公司信息通信分公司 | Power communication equipment state prediction method and system |
| CN111476408B (en) * | 2020-03-26 | 2023-07-25 | 国网江苏省电力有限公司信息通信分公司 | Power communication equipment state prediction method and system |
| CN111401473A (en) * | 2020-04-09 | 2020-07-10 | 中国人民解放军国防科技大学 | Infrared target classification method based on attention mechanism convolutional neural network |
| CN111522956A (en) * | 2020-05-08 | 2020-08-11 | 河南理工大学 | Text emotion classification method based on double channels and hierarchical attention network |
| CN111651593A (en) * | 2020-05-08 | 2020-09-11 | 河南理工大学 | Text emotion analysis method based on word vector and word vector mixed model |
| CN111651603A (en) * | 2020-06-04 | 2020-09-11 | 上海电力大学 | Power industry single-level text classification method and system based on LAV parameter fine adjustment |
| CN111651603B (en) * | 2020-06-04 | 2024-03-19 | 上海电力大学 | Electric power work order hierarchical text classification method and system based on LAV parameter fine adjustment |
| CN111696624B (en) * | 2020-06-08 | 2022-07-12 | 天津大学 | DNA binding protein identification and function annotation deep learning method based on self-attention mechanism |
| CN111696624A (en) * | 2020-06-08 | 2020-09-22 | 天津大学 | DNA binding protein identification and function annotation deep learning method based on self-attention mechanism |
| CN111881262A (en) * | 2020-08-06 | 2020-11-03 | 重庆邮电大学 | Text emotion analysis method based on multi-channel neural network |
| CN111881262B (en) * | 2020-08-06 | 2022-05-20 | 重庆邮电大学 | Text emotion analysis method based on multi-channel neural network |
| CN112818124A (en) * | 2021-02-21 | 2021-05-18 | 昆明理工大学 | Entity relationship extraction method based on attention neural network |
| CN113191138B (en) * | 2021-05-14 | 2023-10-31 | 长江大学 | Automatic text emotion analysis method based on AM-CNN algorithm |
| CN113191138A (en) * | 2021-05-14 | 2021-07-30 | 长江大学 | Automatic text emotion analysis method based on AM-CNN algorithm |
| CN113326374A (en) * | 2021-05-25 | 2021-08-31 | 成都信息工程大学 | Short text emotion classification method and system based on feature enhancement |
| CN114022909A (en) * | 2022-01-07 | 2022-02-08 | 首都师范大学 | Emotion recognition method and system based on sensor data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109299268A (en) | A kind of text emotion analysis method based on dual channel model | |
| CN108763216A (en) | A kind of text emotion analysis method based on Chinese data collection | |
| CN108038205B (en) | Viewpoint analysis prototype system for Chinese microblogs | |
| CN109829166B (en) | People and host customer opinion mining method based on character-level convolutional neural network | |
| CN108984530A (en) | A kind of detection method and detection system of network sensitive content | |
| CN110083833B (en) | Method for analyzing emotion by jointly embedding Chinese word vector and aspect word vector | |
| CN107092596A (en) | Text emotion analysis method based on attention CNNs and CCR | |
| CN103646088B (en) | Product comment fine-grained emotional element extraction method based on CRFs and SVM | |
| CN109376251A (en) | A kind of microblogging Chinese sentiment dictionary construction method based on term vector learning model | |
| CN112199608B (en) | Social media rumor detection method based on network information propagation graph modeling | |
| CN109740154A (en) | A kind of online comment fine granularity sentiment analysis method based on multi-task learning | |
| CN111061843A (en) | Knowledge graph guided false news detection method | |
| CN108984523A (en) | A kind of comment on commodity sentiment analysis method based on deep learning model | |
| CN110245229A (en) | A kind of deep learning theme sensibility classification method based on data enhancing | |
| CN107391760A (en) | User interest recognition methods, device and computer-readable recording medium | |
| CN105512687A (en) | Emotion classification model training and textual emotion polarity analysis method and system | |
| CN110929034A (en) | Commodity comment fine-grained emotion classification method based on improved LSTM | |
| CN106096004A (en) | A kind of method setting up extensive cross-domain texts emotional orientation analysis framework | |
| Bu et al. | A hybrid deep learning system of CNN and LRCN to detect cyberbullying from SNS comments | |
| CN110415071B (en) | Automobile competitive product comparison method based on viewpoint mining analysis | |
| CN110598219A (en) | Emotion analysis method for broad-bean-net movie comment | |
| CN108345587A (en) | A kind of the authenticity detection method and system of comment | |
| CN106778878A (en) | A kind of character relation sorting technique and device | |
| CN108108468A (en) | A kind of short text sentiment analysis method and apparatus based on concept and text emotion | |
| CN111309910A (en) | Text information mining method and device |
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: 20190201 |
|
| RJ01 | Rejection of invention patent application after publication |