CN108205522A - The method and its system of Emotion tagging - Google Patents
The method and its system of Emotion tagging Download PDFInfo
- Publication number
- CN108205522A CN108205522A CN201611173944.4A CN201611173944A CN108205522A CN 108205522 A CN108205522 A CN 108205522A CN 201611173944 A CN201611173944 A CN 201611173944A CN 108205522 A CN108205522 A CN 108205522A
- Authority
- CN
- China
- Prior art keywords
- text
- training
- emotion
- marked
- term vector
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- 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
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/60—Information retrieval; Database structures therefor; File system structures therefor of audio data
- G06F16/68—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/683—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/685—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using automatically derived transcript of audio data, e.g. lyrics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Abstract
The present invention relates to the method and its system of a kind of Emotion tagging, this method includes：Receive text to be marked；Sentiment analysis is carried out to text to be marked using advance trained at least one training pattern, determines the affective tag of text to be marked；For the text marking affective tag to be marked.Using Emotion tagging method provided in an embodiment of the present invention, it is extracted with reference to emotion word and emotion word is converted into word feature vector and carry out model training, and Emotion tagging is carried out to text to be identified, optimize emotional expression ability, meanwhile improve accuracy of the training pattern to sentiment analysis.
Description
Technical field
The method and its system of design data analysis technical field of the present invention more particularly to a kind of Emotion tagging.
Background technology
Lyrics emotional semantic classification typically uses word and word frequency as feature vector, then normal using vector characteristics model etc.
Machine learning is analyzed.But the word and word frequency that usually used song sensibility classification method uses are as feature vector
Granularity it is thicker, precision is not high enough in other words, easily ignores some implicit contacts, such as " I am unhappy " in song,
Using the term vector of coarseness, the word that may be extracted is " happy ", the phenomenon that causing to express mistake or emotional expression reduction,
So that the accuracy rate of emotional semantic classification reduces.
Invention content
The present invention provides the method and its system of a kind of Emotion tagging, the word with reference to emotion word extraction and word2wec is special
The construction method of vector is levied, the ability to express of feature vector is optimized, optimizes text, such as the sentiment analysis effect of the lyrics,
Improve the accuracy rate of text emotion classification.
In a first aspect, an embodiment of the present invention provides a kind of Emotion tagging method, this method can include：
Receive text to be marked；
Sentiment analysis is carried out to text to be marked using advance trained at least one training pattern, determines text to be marked
This affective tag；
For text marking affective tag to be marked.
With reference to emotion word extract and by emotion word be converted to word feature vector carry out model training, and to text to be identified into
Row Emotion tagging optimizes emotional expression ability, meanwhile, improve accuracy rate of the training pattern to sentiment analysis.
Optionally, sentiment analysis is being carried out to text to be marked using advance trained at least one training pattern, really
Before the affective tag of fixed text to be marked, method further includes：
The training set of each affective tag is obtained according at least one affective tag, training set treats training text including multiple
This；
Extract multiple emotion words for treating training text that each training set includes；
Determine the term vector of emotion word；
Model training is carried out to the term vector of multiple emotion words for treating training text that each training set includes, is trained
Model.
Optionally it is determined that the term vector of emotion word, including：
The term vector of emotion word is determined using Word2ved.
Optionally, the term vector of emotion word is determined using Word2ved, including：
The CBOW algorithms included using Word2ved determine the term vector of emotion word.
Optionally, the term vector of emotion word is determined using Word2ved, including：
The Skip-gram algorithms included using Word2ved determine the term vector of emotion word.
Second aspect, an embodiment of the present invention provides a kind of system, which can include：
Receiving unit, for receiving text to be marked；
Processing unit, for carrying out emotion point to text to be marked using advance trained at least one training pattern
Analysis determines the affective tag of text to be marked；
Processing unit is additionally operable to as text marking affective tag to be marked.
With reference to emotion word extract and by emotion word be converted to word feature vector carry out model training, and to text to be identified into
Row Emotion tagging optimizes emotional expression ability, meanwhile, improve accuracy rate of the training pattern to sentiment analysis.
Optionally, system further includes training unit,
Processing unit is additionally operable to obtain the training set of each affective tag, training set packet according at least one affective tag
It includes and multiple treats training text；
Processing unit is additionally operable to extract multiple emotion words for treating training text that each training set includes；
Processing unit is additionally operable to determine the term vector of emotion word；
Training unit, the term vector of multiple emotion words for treating training text for including to each training set carry out model
Training, obtains training pattern.
Optionally, processing unit is specifically used for, and the term vector of emotion word is determined using Word2ved.
Optionally, processing unit is used for,
The CBOW algorithms included using Word2ved determine the term vector of emotion word.
Optionally, processing unit is specifically used for,
The Skip-gram algorithms included using Word2ved determine the term vector of emotion word.
The method and its system of Emotion tagging based on the embodiment of the present invention are extracted with reference to emotion word and by emotion word
It is converted to word feature vector and carries out model training, and Emotion tagging is carried out to text to be identified, optimize emotional expression ability, together
When, improve accuracy rate of the training pattern to sentiment analysis.
Description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, it will make below to required in the embodiment of the present invention
Attached drawing is briefly described, it should be apparent that, drawings described below is only some embodiments of the present invention, for
For those of ordinary skill in the art, without creative efforts, other are can also be obtained according to these attached drawings
Attached drawing.
Fig. 1 is a kind of flow chart of the training method of sentiment classification model provided by the invention；
Fig. 2 is a kind of method flow diagram of Emotion tagging provided in an embodiment of the present invention；
Fig. 3 is a kind of structure diagram of system provided in an embodiment of the present invention.
Specific embodiment
The present invention provides the training method and its system of a kind of sentiment classification model, suitable for the emotion point to text
Class, such as classify to the emotion of the lyrics in lyrics file.
The training method of sentiment classification model provided by the invention is described in detail below in conjunction with the accompanying drawings.
Fig. 1 is a kind of flow chart of the training method of sentiment classification model provided in an embodiment of the present invention.As shown in Figure 1,
This method may comprise steps of：
S110, the training set of each affective tag is obtained according at least one affective tag, and training set is waited to instruct including multiple
Practice text.
Using popular Thayer emotion models, i.e., respectively from two reference axis of energy and pressure by audio data, such as
The emotional semantic classification of song divides two classes, such as the emotional semantic classification of song is formed " vivifying ", and " gratifying " " prevents
Funeral ", " anxiety ", etc. a variety of emotional semantic classifications are divided into a variety of affective tags.
According to affective tag (such as vivifying, satisfactorily, dejected, anxiety waits affective tags) respectively from text
Preset fixed number purpose text file is chosen in this library, forms training set.Such as presetting number is extracted from lyrics library, such as
1000 lyrics file.
It should be noted that presetting number can be configured according to demand.For example carry out emotion according to training pattern
The accurate precision of classification is determined, and usual presetting number is bigger, right when the training pattern that training obtains carries out emotional semantic classification
The accurate precision of the emotional semantic classification of text is higher.
S120 extracts multiple emotion words for treating training text that each training set includes.
The emotion word of extraction text needs to consider a variety of situations, such as influence of the emotion related term to sentiment analysis；Emotion
Word ambiguity present in practical significance, in other words, emotion word ambiguity caused by semantic context；And it negative word and repaiies
Emotion caused by emotion word is put in anti-, emotion enhancing or the effect weakened in excuse, if the lyrics in song are " in high buildings and large mansions
In city, I am unhappy ", negative word " no " causes emotion to emotion word " happiness " and puts reaction.
Based on the above situation of extraction emotion word, segmenter of increasing income may be used in embodiments of the present invention, such as：Breathe out work
Big language platform LTP, the ictclas segmenter of the Chinese Academy of Sciences, SCWS segmenter, segmenter of dismembering an ox as skillfully as a butcher etc. are increased income in segmenter
Any one or multiple segmenter of increasing income are treated to obtain emotion word in training text from multiple.In embodiments of the present invention, it uses
It increases income segmenter, such as LTP carries out the extraction of emotion word so that lyrics sentiment analysis is more convenient, accurate.
It will be extracted according to multiple emotion words treated in training text that affective tag extracts, and by each feelings of extraction
Feel the corresponding emotion word generation emotion list of label, each emotion word in emotion list includes an emotion intensity value, emotion
Intensity value may range from 0 to 1.
Such as：The lyrics " in the city of high buildings and large mansions, I am unhappy " in song, the emotion word of extraction can be " no "
" happiness ", after logic decision, emotion intensity value is " happiness 0 ", wherein, 0 represents that emotion intensity value is minimum, and 1 represents feelings
Feel intensity value highest.Happiness 0 represents that happiness value is minimum.
It should be noted that in embodiments of the present invention, emotion intensity value is set as 0 to 1 value just for the sake of saying
Bright technical scheme of the present invention, in embodiments of the present invention, emotion intensity value may be arranged as the numerical value of other ranges, at this
This is not restricted in inventive embodiments, can self-defined setting be carried out with demand.
S130 determines the term vector of emotion word.
In embodiments of the present invention, word2vec can be used from the term vector of above-mentioned emotion word list generation text.
Word2vec can while language ambience information is captured compressed data scale.
Actually there are two types of different algorithms by word2vec：Continous Bag of Words (CBOW) algorithms and
Skip-gram algorithms.The purpose of CBOW algorithms is that the probability of current term is predicted according to the context of text.Skip-gram
Algorithm is just opposite：The probability of context can be predicted according to current term.Both algorithms all utilize artificial neural network
As their sorting algorithm.Originally, each emotion word is a random N-dimensional vector.By after training, being calculated using CBOW
Method or Skip-gram algorithms obtain the optimal term vector of each emotion word.
Emotion word is extracted and is combined with word2vec so that the term vector built has better emotional expression energy
Power.
S140 carries out model training to the term vector of multiple emotion words for treating training text that each training set includes, obtains
To training pattern.
Model training uses support vector machines (Support Vector Machines, SVM), is have supervision
Model is practised, commonly used to carry out pattern-recognition, classification and regression analysis.The application carries out model instruction using common SVM
Practice, obtain training pattern.
The training pattern that the training method of the sentiment classification model provided through the embodiment of the present invention obtains can improve pair
The accuracy rate of text emotion classification.
It should be noted that the training pattern that the sentiment classification model training method provided through the embodiment of the present invention obtains
80% can be reached to the accuracy rate of text emotion classification, compared with the prior art in pass through the carry out emotion of the term vector of coarseness
Classification, improves the accuracy rate classified to text emotion.
The obtained training pattern of training method for the sentiment classification model that Fig. 1 is provided can be used for Fig. 2 to text to be marked
This Emotion tagging.
Fig. 2 is a kind of method flow diagram of Emotion tagging provided in an embodiment of the present invention.As shown in Fig. 2, the process can be with
Include the following steps：
S210 receives text to be marked.
S220 carries out sentiment analysis to text to be marked using advance trained at least one training pattern, determines to treat
Mark the affective tag of text.
S230 is text marking affective tag to be marked.
Text to be marked is the text of pending emotional semantic classification, in other words affective tag to be marked.It is treated when system receives
When marking text, sentiment analysis is carried out, and determine to treat to text to be marked using advance trained at least one training pattern
The affective tag of text is marked, for example the affective tag of the text to be marked belongs to " vivifying ", " dejected " " makes us
Meet " or " anxiety " etc..
In embodiments of the present invention, trained at least one training pattern can be according to affective tag in Fig. 1 in advance
At least one training pattern that classification based training obtains.
After the affective tag for determining text to be marked, it is text marking affective tag to be marked, completes text to be marked
Emotional semantic classification improves the accuracy rate of the emotional semantic classification of text to be marked.
The training method of sentiment classification model and the process of Emotion tagging is described in detail in figure 1 above and Fig. 2, ties below
Attached drawing 3 is closed system provided in an embodiment of the present invention is described in detail.
Fig. 3 is system provided in an embodiment of the present invention.As shown in figure 3, the system can include receiving unit 310, analysis
Unit 320 and processing unit 330.It should be noted that in embodiments of the present invention, analytic unit 320 and processing unit 330 can
Think processor.
Receiving unit 310, for receiving text to be marked.
Analytic unit 320, for carrying out emotion to text to be marked using advance trained at least one training pattern
Analysis determines the affective tag of text to be marked.
Processing unit 330 is additionally operable to as text marking affective tag to be marked.
System provided in an embodiment of the present invention carries out emotional semantic classification to text to be marked, and carries out the specific of Emotion tagging and retouch
The process of stating can be found in the S220 of Fig. 2, for succinct description, repeat no more herein.
The method and its system of Emotion tagging based on the embodiment of the present invention are extracted with reference to emotion word and by emotion word
It is converted to word feature vector and carries out model training, and Emotion tagging is carried out to text to be identified, emotional expression ability is optimized,
Meanwhile improve accuracy rate of the training pattern to sentiment analysis.
In embodiments of the present invention, as shown in figure 3, the system further includes training unit 340.
Processing unit 330, for obtaining the training set of each affective tag, training set packet according at least one affective tag
It includes and multiple treats training text.
Processing unit unit 330, for extracting multiple emotion words for treating training text that each training set includes.
Processing unit 330, for determining the term vector of emotion word.
Training unit 340, the term vector of multiple emotion words for treating training text for including to each training set carry out
Model training obtains training pattern.
The training pattern that the training method of the sentiment classification model provided through the embodiment of the present invention obtains can improve pair
The accuracy rate of text emotion classification.
Optionally, processing unit 330 is specifically used for, and by increasing income, segmenter is treated to extract emotion in training text from multiple
Word.
Segmenter of increasing income can include Harbin Institute of Technology language platform LTP, the ictclas segmenter of the Chinese Academy of Sciences, SCWS segmenter,
It dismembers an ox as skillfully as a butcher segmenter, etc..In embodiments of the present invention, it LTP may be used treats the emotion word of training text and extract.
Optionally, in embodiments of the present invention, processing unit 330 is specifically used for, and emotion word is determined using Word2ved
Term vector.
In embodiments of the present invention, actually there are two types of different algorithms by word2vec：Continous Bag of
Words (CBOW) algorithms and Skip-gram algorithms.The purpose of CBOW algorithms is to predict current term according to the context of text
Probability.Skip-gram algorithms are just opposite：The probability of context can be predicted according to current term.Both algorithms are all
By the use of artificial neural network as their sorting algorithm.Originally, each emotion word is a random N-dimensional vector.By instruction
After white silk, the optimal term vector of each emotion word is obtained using CBOW algorithms or Skip-gram algorithms.
Optionally, in embodiments of the present invention, processing unit 330 is used for, and the CBOW algorithms included using Word2ved are true
Determine the term vector of emotion word.
Optionally, in embodiments of the present invention, processing unit 330 is specifically used for, the Skip- included using Word2ved
Gram algorithms determine the term vector of emotion word.
Emotion word is extracted and is combined with word2vec so that the term vector built has better emotional expression energy
Power.
Each unit in the system that Fig. 3 is provided can complete method/step S110, S120, S130 in Fig. 1 and
S210, S220 and S230 of S140 and Fig. 2, for succinct description, details are not described herein.
The system provided through the embodiment of the present invention can improve the accuracy of the emotional semantic classification of text to be marked.
Above-described specific embodiment has carried out the purpose of the present invention, technical solution and advantageous effect further
It is described in detail, it should be understood that the foregoing is merely the specific embodiment of the present invention, is not intended to limit the present invention
Protection domain, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should all include
Within protection scope of the present invention.
Claims (10)
- A kind of 1. method of Emotion tagging, which is characterized in that the method includes：Receive text to be marked；Sentiment analysis is carried out to the text to be marked using advance trained at least one training pattern, determines described to wait to mark The affective tag of explanatory notes sheet；For affective tag described in the text marking to be marked.
- 2. according to the method described in claim 1, it is characterized in that, described using advance trained at least one trained mould Type is to the text progress sentiment analysis to be marked, and before the affective tag for determining the text to be marked, the method is also wrapped It includes：The training set of each affective tag is obtained according at least one affective tag, the training set treats training text including multiple This；The multiple emotion word for treating training text that each training set of extraction includes；Determine the term vector of the emotion word；Model training is carried out to the term vector of multiple emotion words for treating training text that each training set includes, is obtained Training pattern.
- 3. method according to claim 1 or 2, which is characterized in that the term vector for determining the emotion word, including：The term vector of the emotion word is determined using Word2ved.
- 4. according to the method described in right 3, which is characterized in that the term vector that the emotion word is determined using Word2ved, Including：The CBOW algorithms included using the Word2ved determine the term vector of the emotion word.
- 5. according to the method described in claim 3, it is characterized in that, it is described using Word2ved determine the word of the emotion word to Amount, including：The Skip-gram algorithms included using the Word2ved determine the term vector of the emotion word.
- 6. a kind of system, which is characterized in that the system comprises：Receiving unit, for receiving text to be marked；Processing unit, for carrying out emotion point to the text to be marked using advance trained at least one training pattern Analysis determines the affective tag of the text to be marked；The processing unit is additionally operable to as affective tag described in the text marking to be marked.
- 7. system according to claim 6, which is characterized in that the system also includes training unit,The processing unit is additionally operable to obtain the training set of each affective tag, the training according at least one affective tag Collection includes multiple treating training text；The processing unit is additionally operable to the multiple emotion word for treating training text that each training set of extraction includes；The processing unit is additionally operable to determine the term vector of the emotion word；The training unit, for the term vector of multiple emotion words for treating training text included to each training set Model training is carried out, obtains training pattern.
- 8. the system described according to claim 6 or 7, which is characterized in that the processing unit is specifically used for, and uses Word2ved Determine the term vector of the emotion word.
- 9. according to the system described in right 8, which is characterized in that the processing unit is used for,The CBOW algorithms included using the Word2ved determine the term vector of the emotion word.
- 10. system according to claim 8, which is characterized in that the processing unit is specifically used for,The Skip-gram algorithms included using the Word2ved determine the term vector of the emotion word.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611173944.4A CN108205522A (en) | 2016-12-16 | 2016-12-16 | The method and its system of Emotion tagging |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611173944.4A CN108205522A (en) | 2016-12-16 | 2016-12-16 | The method and its system of Emotion tagging |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108205522A true CN108205522A (en) | 2018-06-26 |
Family
ID=62602724
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611173944.4A Pending CN108205522A (en) | 2016-12-16 | 2016-12-16 | The method and its system of Emotion tagging |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108205522A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109933686A (en) * | 2019-03-18 | 2019-06-25 | 阿里巴巴集团控股有限公司 | Song Tag Estimation method, apparatus, server and storage medium |
| CN110704581A (en) * | 2019-09-11 | 2020-01-17 | 阿里巴巴集团控股有限公司 | Computer-executed text emotion analysis method and device |
| CN111582360A (en) * | 2020-05-06 | 2020-08-25 | 北京字节跳动网络技术有限公司 | Method, apparatus, device and medium for labeling data |
| CN111973178A (en) * | 2020-08-14 | 2020-11-24 | 中国科学院上海微系统与信息技术研究所 | Electroencephalogram signal identification system and method |
| WO2021052137A1 (en) * | 2019-09-20 | 2021-03-25 | 平安科技（深圳）有限公司 | Emotion vector generation method and apparatus |
| CN117172248A (en) * | 2023-11-03 | 2023-12-05 | 翼方健数(北京)信息科技有限公司 | Text data labeling method, system and medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050228643A1 (en) * | 2004-03-23 | 2005-10-13 | Munteanu Dragos S | Discovery of parallel text portions in comparable collections of corpora and training using comparable texts |
| CN105243129A (en) * | 2015-09-30 | 2016-01-13 | 清华大学深圳研究生院 | Commodity property characteristic word clustering method |
| CN105740349A (en) * | 2016-01-25 | 2016-07-06 | 重庆邮电大学 | Sentiment classification method capable of combining Doc2vce with convolutional neural network |
| CN105824922A (en) * | 2016-03-16 | 2016-08-03 | 重庆邮电大学 | Emotion classifying method fusing intrinsic feature and shallow feature |
| CN105930368A (en) * | 2016-04-13 | 2016-09-07 | 深圳大学 | Emotion classification method and system |
-
2016
- 2016-12-16 CN CN201611173944.4A patent/CN108205522A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050228643A1 (en) * | 2004-03-23 | 2005-10-13 | Munteanu Dragos S | Discovery of parallel text portions in comparable collections of corpora and training using comparable texts |
| CN105243129A (en) * | 2015-09-30 | 2016-01-13 | 清华大学深圳研究生院 | Commodity property characteristic word clustering method |
| CN105740349A (en) * | 2016-01-25 | 2016-07-06 | 重庆邮电大学 | Sentiment classification method capable of combining Doc2vce with convolutional neural network |
| CN105824922A (en) * | 2016-03-16 | 2016-08-03 | 重庆邮电大学 | Emotion classifying method fusing intrinsic feature and shallow feature |
| CN105930368A (en) * | 2016-04-13 | 2016-09-07 | 深圳大学 | Emotion classification method and system |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109933686A (en) * | 2019-03-18 | 2019-06-25 | 阿里巴巴集团控股有限公司 | Song Tag Estimation method, apparatus, server and storage medium |
| CN109933686B (en) * | 2019-03-18 | 2023-02-03 | 创新先进技术有限公司 | Song label prediction method, device, server and storage medium |
| CN110704581A (en) * | 2019-09-11 | 2020-01-17 | 阿里巴巴集团控股有限公司 | Computer-executed text emotion analysis method and device |
| CN110704581B (en) * | 2019-09-11 | 2024-03-08 | 创新先进技术有限公司 | Text emotion analysis method and device executed by computer |
| WO2021052137A1 (en) * | 2019-09-20 | 2021-03-25 | 平安科技（深圳）有限公司 | Emotion vector generation method and apparatus |
| CN111582360A (en) * | 2020-05-06 | 2020-08-25 | 北京字节跳动网络技术有限公司 | Method, apparatus, device and medium for labeling data |
| CN111582360B (en) * | 2020-05-06 | 2023-08-15 | 北京字节跳动网络技术有限公司 | Method, apparatus, device and medium for labeling data |
| CN111973178A (en) * | 2020-08-14 | 2020-11-24 | 中国科学院上海微系统与信息技术研究所 | Electroencephalogram signal identification system and method |
| CN117172248A (en) * | 2023-11-03 | 2023-12-05 | 翼方健数(北京)信息科技有限公司 | Text data labeling method, system and medium |
| CN117172248B (en) * | 2023-11-03 | 2024-01-30 | 翼方健数(北京)信息科技有限公司 | Text data labeling method, system and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108205522A (en) | The method and its system of Emotion tagging | |
| Latif et al. | Transfer learning for improving speech emotion classification accuracy | |
| CN109918673B (en) | Semantic arbitration method and device, electronic equipment and computer-readable storage medium | |
| Zeng et al. | Distant supervision for relation extraction via piecewise convolutional neural networks | |
| CN107291783B (en) | Semantic matching method and intelligent equipment | |
| CN105260356B (en) | Chinese interaction text emotion and topic detection method based on multi-task learning | |
| CN108255805A (en) | The analysis of public opinion method and device, storage medium, electronic equipment | |
| Latif et al. | Cross corpus speech emotion classification-an effective transfer learning technique | |
| CN111241232B (en) | Business service processing method and device, service platform and storage medium | |
| CN112507704B (en) | Multi-intention recognition method, device, equipment and storage medium | |
| CN108304373A (en) | Construction method, device, storage medium and the electronic device of semantic dictionary | |
| CN108899033A (en) | A kind of method and device of determining speaker characteristic | |
| CN108345612A (en) | A kind of question processing method and device, a kind of device for issue handling | |
| CN111159332A (en) | Text multi-intention identification method based on bert | |
| CN110287311A (en) | File classification method and device, storage medium, computer equipment | |
| de Velasco et al. | Emotion Detection from Speech and Text. | |
| CN111475651A (en) | Text classification method, computing device and computer storage medium | |
| CN113609865A (en) | Text emotion recognition method and device, electronic equipment and readable storage medium | |
| Suzuki et al. | Fine-grained named entity classification with wikipedia article vectors | |
| dos Santos | Think positive: Towards Twitter sentiment analysis from scratch | |
| CN109885680B (en) | Short text classification preprocessing method, system and device based on semantic extension | |
| Castellucci et al. | Context-aware convolutional neural networks for twitter sentiment analysis in italian | |
| Fieri et al. | Offensive language detection using soft voting ensemble model | |
| CN116701636A (en) | Data classification method, device, equipment and storage medium | |
| CN106776568A (en) | Based on the rationale for the recommendation generation method that user evaluates |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180626 |