CN112860901A - Emotion analysis method and device integrating emotion dictionaries - Google Patents
Emotion analysis method and device integrating emotion dictionaries Download PDFInfo
- Publication number
- CN112860901A CN112860901A CN202110346235.6A CN202110346235A CN112860901A CN 112860901 A CN112860901 A CN 112860901A CN 202110346235 A CN202110346235 A CN 202110346235A CN 112860901 A CN112860901 A CN 112860901A
- Authority
- CN
- China
- Prior art keywords
- emotion
- sentence
- field
- encoder
- target field
- 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
- 230000008451 emotion Effects 0.000 title claims abstract description 161
- 238000004458 analytical method Methods 0.000 title claims abstract description 36
- 239000013598 vector Substances 0.000 claims abstract description 71
- 238000000034 method Methods 0.000 claims abstract description 29
- 238000012549 training Methods 0.000 claims description 29
- 230000002996 emotional effect Effects 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 13
- 230000015654 memory Effects 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 10
- 238000012512 characterization method Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 6
- 230000000694 effects Effects 0.000 abstract description 2
- 238000013473 artificial intelligence Methods 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 11
- 230000005012 migration Effects 0.000 description 6
- 238000013508 migration Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000013136 deep learning model Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 108010001267 Protein Subunits Proteins 0.000 description 1
- 239000007795 chemical reaction product Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—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
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/374—Thesaurus
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Machine Translation (AREA)
Abstract
The invention provides an emotion analysis method and device fused with an emotion dictionary, which can be applied to the field of artificial intelligence and comprises the following steps: inputting the sentences in the source field and the sentences in the target field into a pre-generated encoder to obtain sentence word vectors of the source field and the sentences of the target field; inputting the source field word vector and the target field sentence word vector into a sentence vector encoder to respectively obtain a source field sentence representation and a target field sentence representation; and obtaining the emotion classification result of the target field by utilizing the source field sentence representation and the target field sentence representation. According to the method and the device, emotion words are used as prediction targets, the linguistic data of the source field and the linguistic data of the target field are jointly trained through the language model, an encoder capable of extracting emotion analysis related features is generated, then a word encoder is used for generating feature vectors of texts to be subjected to emotion classification, and the technical effect that usable classification results can be obtained by using less or no labeled data is achieved.
Description
Technical Field
The application belongs to the technical field of natural language processing, and particularly relates to an emotion analysis method and device integrating emotion dictionaries.
Background
With the wide application of deep learning models in the field of natural language processing, the accuracy of text emotion analysis results is greatly improved, and at present, deep learning models are used for analyzing commodity comments, public opinions and the like. In the financial industry, the emotional index of the public to the legal person is judged by effectively analyzing the current affair news related to the legal person client, and the method has great significance for credit wind control of banks. The emotion analysis aiming at the public opinion of the legal client can provide assistant decision support for the bank to monitor the risk condition of the legal client. However, most of the open source annotation data for emotion classification come from commodity comments and service comments of internet companies at present, and training sample data for news in the financial field is less, so that higher requirements are put forward on an emotion analysis model of financial news, and in order to obtain a better emotion classification result, partial data needs to be manually annotated as a training data set, which causes waste of human resources.
Disclosure of Invention
The application provides an emotion analysis method and device fused with an emotion dictionary, and aims to at least solve the problem of how to obtain emotion classification results without labeling data under the condition that the amount of open source labeled data is small.
According to one aspect of the application, an emotion analysis method fused with an emotion dictionary is provided, and comprises the following steps:
inputting the sentences in the source field and the sentences in the target field into a pre-generated encoder to obtain sentence word vectors of the source field and the sentences of the target field;
inputting the source field word vector and the target field sentence word vector into a sentence vector encoder to respectively obtain a source field sentence representation and a target field sentence representation;
and obtaining the emotion classification result of the target field by utilizing the source field sentence representation and the target field sentence representation.
In one embodiment, obtaining emotion classification results of a target domain by using source domain sentence characterization and target domain sentence characterization comprises:
training the emotion classifier by using the source field sentence representation with the label to obtain the trained emotion classifier;
and carrying out emotion classification on the annotation-free target field sentence representation through the trained emotion classifier to obtain an emotion classification result of the target field.
In one embodiment, the sentence vector encoder is a long-short term memory network encoder.
In one embodiment, the emotion classifier is an MLP multi-layer perceptron.
In one embodiment, the generation process of the encoder includes:
acquiring an emotion dictionary of an emotion dictionary target field of a source field;
fuzzy processing is carried out on emotional words in an emotional dictionary appearing in the sentence;
and training a word vector by using the fuzzy sentences through a language model to obtain an encoder capable of extracting emotional characteristics.
In an embodiment, the emotion analysis method fusing the emotion dictionaries further includes:
and optimizing the encoder by using an Adam optimizer to obtain an optimized encoder.
According to another aspect of the present application, there is also provided an emotion analysis apparatus fusing an emotion dictionary, including:
a word vector acquiring unit, configured to input the acquired sentences in the source domain and the target domain into a pre-generated encoder to acquire a source domain sentence word vector and a target domain sentence word vector;
a sentence representation obtaining unit, configured to input the source field word vector and the target field sentence word vector into a sentence vector encoder to obtain a source field sentence representation and a target field sentence representation, respectively;
and the emotion classification unit is used for obtaining an emotion classification result of the target field by utilizing the source field sentence representation and the target field sentence representation.
In one embodiment, the emotion classification unit includes:
the emotion classifier training module is used for training the emotion classifier by using the source field sentence representation with the label to obtain the trained emotion classifier;
and the classification module is used for carrying out emotion classification on the sentence representations of the target field without the labels through the trained emotion classifier to obtain the emotion classification result of the target field.
In one embodiment, the sentence vector encoder is a long-short term memory network encoder.
In one embodiment, the emotion classifier is an MLP multi-layer perceptron.
In one embodiment, the emotion analyzing apparatus further has a generating unit of an encoder, including:
the emotion dictionary acquisition module is used for acquiring an emotion dictionary of an emotion dictionary target field in a source field;
the fuzzy module is used for carrying out fuzzy processing on the emotional words in the emotional dictionary appearing in the sentence;
and the training module is used for training the word vector by using the fuzzy sentences through the language model to obtain the encoder capable of extracting the emotional characteristics.
In one embodiment, the emotion analyzing apparatus further includes:
and the optimization module is used for optimizing the encoder by using an Adam optimizer to obtain the optimized encoder.
According to the method, the emotion words are used as prediction targets, the linguistic data of the source field and the linguistic data of the target field are subjected to combined training through the language model, an encoder capable of extracting emotion analysis related features is generated, then the word encoder is used for generating feature vectors of texts to be subjected to emotion classification, and the technical effect that available classification results can be obtained with little or no label data is achieved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of an emotion analysis method fusing an emotion dictionary provided by the present application.
FIG. 2 is a flowchart of a method for obtaining emotion classification results in a target domain in an embodiment of the present application.
Fig. 3 is a flowchart of a generation process of an encoder in an embodiment of the present application.
Fig. 4 is a schematic diagram of a word vector generation model in the embodiment of the present application.
Fig. 5 is a block diagram illustrating a structure of an emotion analyzing apparatus incorporating an emotion dictionary according to the present application.
FIG. 6 is a block diagram of the emotion classification unit in the embodiment of the present application.
Fig. 7 is a block diagram of a generating unit of an encoder in the embodiment of the present application.
Fig. 8 is a specific implementation of an electronic device in an embodiment of the present application.
Fig. 9 is a structural diagram of an overall framework of migration learning in the embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Most open source annotation data aiming at emotion classification come from commodity comments and service comments of internet companies at present, training sample data aiming at news in the financial field is less, so that higher requirements are put forward on an emotion analysis model of the financial news, and in order to obtain a better emotion classification result, partial data needs to be artificially annotated as a training data set, so that waste of human resources is caused. Therefore, in order to fully utilize labeled data in other existing fields to solve the problem of data certainty in the field, a migration learning mode can be adopted to enable the model to migrate from a source field with labeled data to a target field without labeled data, so as to train the model suitable for the target field.
Based on this, the present application provides an emotion analysis method fusing an emotion dictionary, as shown in fig. 1, including:
s101: and inputting the obtained sentences in the source field and the sentences in the target field into a pre-generated encoder to obtain word vectors of the sentences in the source field and the sentences in the target field.
S102: and inputting the source field word vector and the target field sentence word vector into a sentence vector encoder to respectively obtain a source field sentence representation and a target field sentence representation.
S103: and obtaining the emotion classification result of the target field by utilizing the source field sentence representation and the target field sentence representation.
The source domain is other domains with labeled data in the prior art, and the target domain is a domain without labeled data to be subjected to emotion classification. Firstly, before emotion classification is performed on a text in a target field, a word vector encoder (described in detail later) which integrates knowledge of a source field and knowledge of the target field is generated, sentences in the source field and sentences in the target field are processed by the word vector encoder to obtain word vectors, the word vectors are input into the sentence vector encoder to generate sentence representations, and then the sentence representations in the source field are used for classifier training to finally obtain emotion classification in the target field.
In one embodiment, obtaining emotion classification results of a target domain by using a source domain sentence characterization and a target domain sentence characterization, as shown in fig. 2, includes:
s201: and training the emotion classifier by using the source field sentence representation with the label to obtain the trained emotion classifier.
S202: and carrying out emotion classification on the annotation-free target field sentence representation through the trained emotion classifier to obtain an emotion classification result of the target field.
In a specific embodiment, after a sentence representation is generated by an LSTM (Long Short-Term Memory network) according to a pre-generated encoder with an obtained source domain and target domain word vector representation, the sentence representation is transferred to an annotation-free target domain emotion prediction after an emotion classifier is trained by annotated metadata, which specifically includes: firstly, LSTM is used as a sentence vector encoder, sentences in a source field and a target field text are processed by a pre-generated encoder to obtain word vectors, and the word vectors are input into the LSTM to generate sentence representations (a source field sentence representation S and a target field sentence representation T);
and then, inputting the sentence characterization S and the label Ls of the source field with the labeled data into a classifier to perform learning training on the classifier, wherein in the application, the emotion classifier C is obtained by training, and the application is not limited to this.
And finally, inputting the sentence representation T in the target field without the labeled data into the learned emotion classifier C to obtain the emotion classification result of the target field data.
In one embodiment, the sentence vector encoder is a long-short term memory network encoder.
In one embodiment, the emotion classifier is an MLP multi-layer perceptron.
In one embodiment, as shown in fig. 3, the generation process of the encoder includes:
s301: and acquiring the emotion dictionary of the emotion dictionary target field of the source field.
S302: and carrying out fuzzy processing on the emotional words in the emotional dictionary appearing in the sentence.
S303: and training a word vector by using the fuzzy sentences through a language model to obtain an encoder capable of extracting emotional characteristics.
In an embodiment, the emotion analysis method fusing the emotion dictionaries further includes:
and optimizing the encoder by using an Adam optimizer to obtain an optimized encoder.
In a specific embodiment, according to emotion dictionaries of a source field and a target field, emotion words in original linguistic data of the two fields are respectively subjected to mask (fuzzy); and (3) carrying out unsupervised training by taking what the next word of the language model is as a target, adopting a transformer as an encoder to train a word vector generation model for the semantic representation that the mask word simultaneously has context, wherein the word vector generation model is shown as figure 4, and the overall framework of the migration learning is shown as figure 9. And inputting the fuzzy linguistic data into the word vector generation model to predict what the words are of the mask as a training target and outputting a training result, wherein the training result comprises a source field and a target field, and finally training a transformer encoder capable of extracting wider emotional characteristics. Then, the Adam optimizer is used to optimize the encoder to obtain a final encoder. After the sentences in the source domain or the target domain are input into a final encoder, word vector representations related to the emotional features can be generated.
The migration learning method integrating the financial emotion dictionary information can reduce manual labeling dependence of data, saves manpower resources, directly uses emotion words in the financial emotion dictionary words as supervision information, extracts words having direct influence on emotion analysis, can explicitly extract emotion analysis features, synchronously trains source field and target field data sets through a Transformer encoder, is more sufficient in semantic relation between the source field and the target field, and can improve the accuracy of migration learning to a certain extent.
Based on the same inventive concept, the embodiment of the present application further provides an emotion analysis device fused with an emotion dictionary, which can be used for implementing the method described in the above embodiment, as described in the following embodiment. Because the problem solving principle of the emotion analysis device integrated with the emotion dictionary is similar to that of the emotion analysis method integrated with the emotion dictionary, the implementation of the emotion analysis device integrated with the emotion dictionary can refer to the implementation of the emotion analysis method integrated with the emotion dictionary, and repeated details are omitted. As used hereinafter, the term "unit" or "module" may be a combination of software and/or hardware that implements a predetermined function. While the system described in the embodiments below is preferably implemented in software, implementations in hardware, or a combination of software and hardware are also possible and contemplated.
According to another aspect of the present application, there is also provided an emotion analysis apparatus incorporating an emotion dictionary, as shown in fig. 5, including:
a word vector obtaining unit 501, configured to input the obtained sentences in the source domain and the obtained sentences in the target domain into a pre-generated encoder to obtain word vectors of sentences in the source domain and words of sentences in the target domain;
a sentence representation obtaining unit 502, configured to input the source field word vector and the target field sentence word vector into a sentence vector encoder to obtain a source field sentence representation and a target field sentence representation, respectively;
and an emotion classification unit 503, configured to obtain an emotion classification result of the target domain by using the source domain sentence representation and the target domain sentence representation.
In one embodiment, as shown in FIG. 6, the emotion classification unit 503 includes:
the emotion classifier training module 601 is used for training an emotion classifier by using the source field sentence representation with the label to obtain the trained emotion classifier;
the classification module 602 is configured to perform emotion classification on the annotation-free sentence representations in the target field by using the trained emotion classifier to obtain an emotion classification result in the target field.
In one embodiment, the sentence vector encoder is a long-short term memory network encoder.
In one embodiment, the emotion classifier is an MLP multi-layer perceptron.
In one embodiment, as shown in fig. 7, the emotion analyzing apparatus further includes a generating unit of an encoder, including:
an emotion dictionary obtaining module 701, configured to obtain an emotion dictionary of an emotion dictionary target field in a source field;
a fuzzy module 702, configured to perform fuzzy processing on emotion words in an emotion dictionary appearing in a sentence;
and the training module 703 is configured to train a word vector through a language model by using the blurred sentence, so as to obtain an encoder capable of extracting emotional features.
In one embodiment, the emotion analyzing apparatus further includes:
and the optimization module is used for optimizing the encoder by using an Adam optimizer to obtain the optimized encoder.
The application provides a word vector generating device fused with a financial emotion dictionary, which takes emotion words as prediction targets, jointly trains linguistic data of a general field (source field) and a financial field (target field) through a language model, generates an encoder capable of extracting emotion analysis related features, and further generates feature vectors of texts to be subjected to emotion classification by using the encoder; in the migration learning stage, sentence vector representations of a source field are obtained firstly, an emotion classifier is trained through labeled data of the source field, and emotion analysis and judgment are carried out on the sentence representations of a target field through the trained classifier.
An embodiment of the present application further provides a specific implementation manner of an electronic device capable of implementing all steps in the method in the foregoing embodiment, and referring to fig. 8, the electronic device specifically includes the following contents:
a processor (processor)801, a memory 802, a communication Interface 803, a bus 804, and a non-volatile memory 805;
the processor 801, the memory 802 and the communication interface 803 complete mutual communication through the bus 804;
the processor 801 is configured to call the computer programs in the memory 802 and the non-volatile memory 805, and when the processor executes the computer programs, the processor implements all the steps in the method in the foregoing embodiments, for example, when the processor executes the computer programs, the processor implements the following steps:
s101: and inputting the obtained sentences in the source field and the sentences in the target field into a pre-generated encoder to obtain word vectors of the sentences in the source field and the sentences in the target field.
S102: and inputting the source field word vector and the target field sentence word vector into a sentence vector encoder to respectively obtain a source field sentence representation and a target field sentence representation.
S103: and obtaining the emotion classification result of the target field by utilizing the source field sentence representation and the target field sentence representation.
Embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps of the method in the above embodiments, where the computer-readable storage medium stores thereon a computer program, and the computer program when executed by a processor implements all the steps of the method in the above embodiments, for example, the processor implements the following steps when executing the computer program:
s101: and inputting the obtained sentences in the source field and the sentences in the target field into a pre-generated encoder to obtain word vectors of the sentences in the source field and the sentences in the target field.
S102: and inputting the source field word vector and the target field sentence word vector into a sentence vector encoder to respectively obtain a source field sentence representation and a target field sentence representation.
S103: and obtaining the emotion classification result of the target field by utilizing the source field sentence representation and the target field sentence representation.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the hardware + program class embodiment, since it is substantially similar to the method embodiment, the description is simple, and the relevant points can be referred to the partial description of the method embodiment. Although embodiments of the present description provide method steps as described in embodiments or flowcharts, more or fewer steps may be included based on conventional or non-inventive means. The order of steps recited in the embodiments is merely one manner of performing the steps in a multitude of orders and does not represent the only order of execution. When an actual apparatus or end product executes, it may execute sequentially or in parallel (e.g., parallel processors or multi-threaded environments, or even distributed data processing environments) according to the method shown in the embodiment or the figures. The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, the presence of additional identical or equivalent elements in a process, method, article, or apparatus that comprises the recited elements is not excluded. For convenience of description, the above devices are described as being divided into various modules by functions, and are described separately. Of course, in implementing the embodiments of the present description, the functions of each module may be implemented in one or more software and/or hardware, or a module implementing the same function may be implemented by a combination of multiple sub-modules or sub-units, and the like. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form. The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks. As will be appreciated by one skilled in the art, embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present description may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and so forth) having computer-usable program code embodied therein. The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment. In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of an embodiment of the specification. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction. The above description is only an example of the embodiments of the present disclosure, and is not intended to limit the embodiments of the present disclosure. Various modifications and variations to the embodiments described herein will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the embodiments of the present specification should be included in the scope of the claims of the embodiments of the present specification.
Claims (10)
1. An emotion analysis method fused with an emotion dictionary is characterized by comprising the following steps:
inputting the sentences in the source field and the sentences in the target field into a pre-generated encoder to obtain sentence word vectors of the source field and the sentences of the target field;
inputting the source field word vector and the target field sentence word vector into a sentence vector encoder to respectively obtain a source field sentence representation and a target field sentence representation;
and obtaining the emotion classification result of the target field by utilizing the source field sentence representation and the target field sentence representation.
2. The emotion analysis method fused with emotion dictionary according to claim 1, wherein the obtaining of emotion classification result of target domain by using the source domain sentence characterization and the target domain sentence characterization comprises:
training an emotion classifier by using the source field sentence representation with the label to obtain the trained emotion classifier;
and carrying out emotion classification on the target field sentence representations without labels through the trained emotion classifier to obtain an emotion classification result of the target field.
3. The emotion analysis method fused with emotion dictionaries of claim 2, wherein the sentence vector encoder is a long-short term memory network encoder.
4. The emotion analysis method fused with an emotion dictionary according to claim 2, wherein the emotion classifier is an MLP multi-layer perceptron.
5. The emotion analyzing method fused with an emotion dictionary according to claim 1, wherein the generation process of the encoder includes:
acquiring an emotion dictionary of an emotion dictionary target field of a source field;
fuzzy processing is carried out on emotional words in an emotional dictionary appearing in the sentence;
and training a word vector by using the fuzzy sentences through a language model to obtain an encoder capable of extracting emotional characteristics.
6. The emotion analysis method fused with an emotion dictionary according to claim 5, further comprising:
and optimizing the encoder by using an Adam optimizer to obtain an optimized encoder.
7. The emotion analysis method fused with an emotion dictionary as set forth in claim 5, wherein the step of training a word vector by a language model using the blurred sentence comprises:
generating a training sample according to the sentences subjected to the fuzzy processing of the emotional words;
and inputting the training sample into a language model to obtain the meaning of the fuzzy processing emotional words.
8. An emotion analysis device that incorporates an emotion dictionary, comprising:
a word vector acquiring unit, configured to input the acquired sentences in the source domain and the target domain into a pre-generated encoder to acquire a source domain sentence word vector and a target domain sentence word vector;
a sentence representation obtaining unit, configured to input the source field word vector and the target field sentence word vector into a sentence vector encoder to obtain a source field sentence representation and a target field sentence representation, respectively;
and the emotion classification unit is used for obtaining an emotion classification result of the target field by utilizing the source field sentence representation and the target field sentence representation.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the emotion analysis method by fusing emotion dictionaries according to any of claims 1 to 7 when executing the program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the emotion analysis method fusing emotion dictionaries according to any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110346235.6A CN112860901A (en) | 2021-03-31 | 2021-03-31 | Emotion analysis method and device integrating emotion dictionaries |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110346235.6A CN112860901A (en) | 2021-03-31 | 2021-03-31 | Emotion analysis method and device integrating emotion dictionaries |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112860901A true CN112860901A (en) | 2021-05-28 |
Family
ID=75991999
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110346235.6A Pending CN112860901A (en) | 2021-03-31 | 2021-03-31 | Emotion analysis method and device integrating emotion dictionaries |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112860901A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115080688A (en) * | 2022-06-13 | 2022-09-20 | 华南理工大学 | Method and device for analyzing low-sample cross-domain emotion |
| CN116089602A (en) * | 2021-11-04 | 2023-05-09 | 腾讯科技（深圳）有限公司 | Information processing method, apparatus, electronic device, storage medium, and program product |
| CN116778967A (en) * | 2023-08-28 | 2023-09-19 | 清华大学 | Multi-mode emotion recognition method and device based on pre-training model |
| CN117150024A (en) * | 2023-10-27 | 2023-12-01 | 北京电子科技学院 | Cross-domain fine granularity emotion analysis method, system, equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108170681A (en) * | 2018-01-15 | 2018-06-15 | 中南大学 | Text emotion analysis method, system and computer readable storage medium |
| CN109325112A (en) * | 2018-06-27 | 2019-02-12 | 北京大学 | A kind of across language sentiment analysis method and apparatus based on emoji |
| US20190197105A1 (en) * | 2017-12-21 | 2019-06-27 | International Business Machines Corporation | Unsupervised neural based hybrid model for sentiment analysis of web/mobile application using public data sources |
| 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 |
| CN110222178A (en) * | 2019-05-24 | 2019-09-10 | 新华三大数据技术有限公司 | Text sentiment classification method, device, electronic equipment and readable storage medium storing program for executing |
| CN112199505A (en) * | 2020-10-30 | 2021-01-08 | 福州大学 | Cross-domain emotion classification method and system based on feature representation learning |
-
2021
- 2021-03-31 CN CN202110346235.6A patent/CN112860901A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190197105A1 (en) * | 2017-12-21 | 2019-06-27 | International Business Machines Corporation | Unsupervised neural based hybrid model for sentiment analysis of web/mobile application using public data sources |
| CN108170681A (en) * | 2018-01-15 | 2018-06-15 | 中南大学 | Text emotion analysis method, system and computer readable storage medium |
| CN109325112A (en) * | 2018-06-27 | 2019-02-12 | 北京大学 | A kind of across language sentiment analysis method and apparatus based on emoji |
| 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 |
| CN110222178A (en) * | 2019-05-24 | 2019-09-10 | 新华三大数据技术有限公司 | Text sentiment classification method, device, electronic equipment and readable storage medium storing program for executing |
| CN112199505A (en) * | 2020-10-30 | 2021-01-08 | 福州大学 | Cross-domain emotion classification method and system based on feature representation learning |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116089602A (en) * | 2021-11-04 | 2023-05-09 | 腾讯科技（深圳）有限公司 | Information processing method, apparatus, electronic device, storage medium, and program product |
| CN116089602B (en) * | 2021-11-04 | 2024-05-03 | 腾讯科技（深圳）有限公司 | Information processing method, apparatus, electronic device, storage medium, and program product |
| CN115080688A (en) * | 2022-06-13 | 2022-09-20 | 华南理工大学 | Method and device for analyzing low-sample cross-domain emotion |
| CN115080688B (en) * | 2022-06-13 | 2024-06-04 | 华南理工大学 | Cross-domain emotion analysis method and device for few samples |
| CN116778967A (en) * | 2023-08-28 | 2023-09-19 | 清华大学 | Multi-mode emotion recognition method and device based on pre-training model |
| CN116778967B (en) * | 2023-08-28 | 2023-11-28 | 清华大学 | Multi-mode emotion recognition method and device based on pre-training model |
| CN117150024A (en) * | 2023-10-27 | 2023-12-01 | 北京电子科技学院 | Cross-domain fine granularity emotion analysis method, system, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112860901A (en) | Emotion analysis method and device integrating emotion dictionaries | |
| WO2022095682A1 (en) | Text classification model training method, text classification method and apparatus, device, storage medium, and computer program product | |
| JP7150842B2 (en) | Multilingual Document Retrieval Based on Document Structure Extraction | |
| CN110968695A (en) | Intelligent labeling method, device and platform based on active learning of weak supervision technology | |
| WO2021173700A1 (en) | Automated content tagging with latent dirichlet allocation of contextual word embeddings | |
| CN113434685B (en) | Information classification processing method and system | |
| Tang et al. | Research on automatic labeling of imbalanced texts of customer complaints based on text enhancement and layer-by-layer semantic matching | |
| CN112784580A (en) | Financial data analysis method and device based on event extraction | |
| CN111008329A (en) | Page content recommendation method and device based on content classification | |
| Guisiano et al. | SDG-Meter: A deep learning based tool for automatic text classification of the Sustainable Development Goals | |
| Sharma et al. | Various methods to classify the polarity of text based customer reviews using sentiment analysis | |
| CN112926341A (en) | Text data processing method and device | |
| US20230289528A1 (en) | Method for constructing sentiment classification model based on metaphor identification | |
| CN114118068B (en) | Method and device for amplifying training text data and electronic equipment | |
| CN113849634A (en) | Method for improving interpretability of depth model recommendation scheme | |
| Wang et al. | Distill-AER: Fine-Grained Address Entity Recognition from Spoken Dialogue via Knowledge Distillation | |
| CN112860843A (en) | News long text sentiment analysis method and device | |
| Soares et al. | Knowledge driven intelligent survey systems for linguists | |
| Künas et al. | Improving Performance of Long Short-Term Memory Networks for Sentiment Analysis Using Multicore and GPU Architectures | |
| Tung et al. | SemiMemes: A Semi-supervised Learning Approach for Multimodal Memes Analysis | |
| Mnassri et al. | Multilingual Hate Speech Detection Using Semi-supervised Generative Adversarial Network | |
| Sun et al. | Multi-granularity Feature Attention Fusion Network for Image-Text Sentiment Analysis | |
| Hajinasab | A Dynamic Approach to Multi-agent-based Simulation in Urban Transportation Planning | |
| Shilov et al. | Application of Machine Learning Techniques to Enterprise Model Classification: An Approach and First Experimental Results | |
| Bghiel et al. | Visual question answering system for identifying medical images attributes |
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 |