CN110929516A - Text emotion analysis method and device, electronic equipment and readable storage medium - Google Patents
Text emotion analysis method and device, electronic equipment and readable storage medium Download PDFInfo
- Publication number
- CN110929516A CN110929516A CN201911156648.7A CN201911156648A CN110929516A CN 110929516 A CN110929516 A CN 110929516A CN 201911156648 A CN201911156648 A CN 201911156648A CN 110929516 A CN110929516 A CN 110929516A
- Authority
- CN
- China
- Prior art keywords
- emotion
- intensity value
- text
- determining
- processed
- 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 385
- 238000004458 analytical method Methods 0.000 title claims abstract description 77
- 238000000034 method Methods 0.000 claims abstract description 32
- 230000006403 short-term memory Effects 0.000 claims abstract description 7
- 230000002996 emotional effect Effects 0.000 claims description 47
- 230000015654 memory Effects 0.000 claims description 15
- 238000012937 correction Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 4
- 238000012549 training Methods 0.000 description 36
- 239000013598 vector Substances 0.000 description 14
- 238000010586 diagram Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 8
- 238000012360 testing method Methods 0.000 description 7
- 230000011218 segmentation Effects 0.000 description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 238000005034 decoration Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 238000011022 operating instruction Methods 0.000 description 2
- 230000002441 reversible effect Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 235000019800 disodium phosphate Nutrition 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000002787 reinforcement 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
- G06F16/355—Class or cluster creation or modification
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
-
- 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
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Machine Translation (AREA)
Abstract
The embodiment of the application provides a text emotion analysis method and device, electronic equipment and a readable storage medium. The method comprises the following steps: determining a first emotion intensity value of a text to be processed based on a Bi-directional Long Short-Term Memory model; determining a second emotion intensity value of the text to be processed based on a preset emotion expression rule; acquiring a subject term of the text to be processed, and determining a third emotion intensity value of the text to be processed based on the subject term and a preset subject term weight; and determining the emotion analysis result of the text to be processed based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value. According to the scheme, the emotion analysis result of the text to be processed is rapidly acquired through the analysis of the text to be processed, and the emotion polarity of the information issued by the netizens can be acquired in time.
Description
Technical Field
The application relates to the technical field of natural language processing, in particular to a text emotion analysis method and device, an electronic device and a readable storage medium.
Background
With the rise of the popularity of the internet and the mobile internet, medias and the arrival of the age of medias, the information dissemination of the netizens is more and more complicated. In the public opinion propagation of major events, netizens may release a large amount of information, and the emotional polarity of the information released by the netizens cannot be grasped in time through manual work, so how to analyze the large amount of information released by the netizens and acquire the emotional polarity of the netizens in time becomes a problem to be solved urgently in the technical field of natural language processing.
Disclosure of Invention
The present application aims to solve at least one of the above technical drawbacks. The technical scheme adopted by the application is as follows:
in a first aspect, an embodiment of the present application provides a method for emotion analysis of a text, where the method includes:
determining a first emotion intensity value of a text to be processed based on a Bi-directional Long Short-Term Memory model;
determining a second emotion intensity value of the text to be processed based on a preset emotion expression rule;
acquiring a subject term of the text to be processed, and determining a third emotion intensity value of the text to be processed based on the subject term and a preset subject term weight;
and determining the emotion analysis result of the text to be processed based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value.
Optionally, determining a second emotion intensity value of the text to be processed based on a predetermined emotion expression rule includes:
dividing the whole sentence in the text to be processed into clauses according to punctuations in the text to be processed;
determining a fourth emotion intensity value of the clause;
a second sentimental intensity value is determined based on the fourth sentimental intensity value.
Optionally, determining a fourth emotion intensity value of the clause includes:
determining the emotional words, the negative words for modifying the emotional words and the degree adverbs for modifying the emotional words in the clauses;
determining sentence patterns of the clauses;
and determining a fourth emotion intensity value based on the preset weight corresponding to the emotion words, the preset weight coefficient corresponding to the negative words and the preset weight coefficient corresponding to the degree adverbs and the preset weight coefficient corresponding to the sentence patterns.
Optionally, determining the second emotion intensity value based on the fourth emotion intensity value includes:
determining the sentence relation between each clause and the adjacent clauses based on the associated words in each clause in the text to be processed;
and determining a second emotion intensity value based on the fourth emotion intensity value and a preset weight coefficient corresponding to the inter-sentence relation.
Optionally, determining an emotion analysis result of the text to be processed based on the first emotion intensity value, the second emotion intensity value, and the third emotion intensity value includes:
determining a fifth emotion intensity value of the text to be processed based on the first emotion intensity value, a preset weight coefficient corresponding to the first emotion intensity value, the second emotion intensity value, a preset weight coefficient corresponding to the second emotion intensity value, the third emotion intensity value, a preset weight coefficient corresponding to the third emotion intensity value and a preset emotion intensity correction coefficient;
and determining an emotion analysis result of the text to be processed based on the fifth emotion intensity value.
Optionally, the emotion analysis result of the text to be processed includes an emotion polarity of the text to be processed, and determining the emotion analysis result of the text to be processed based on the fifth emotion intensity value includes:
and determining the emotion polarity of the text to be processed based on the fifth emotion intensity value and a preset emotion intensity threshold value.
In a second aspect, an embodiment of the present application provides an emotion analysis apparatus for a text, where the apparatus includes:
the first emotional intensity determining module is used for determining a first emotional intensity value of the text to be processed based on the Bi-directional Long Short-Term Memory model;
the second emotion intensity determination module is used for determining a second emotion intensity value of the text to be processed based on a preset emotion expression rule;
the third emotional intensity determination module is used for acquiring the subject term of the text to be processed and determining a third emotional intensity value of the text to be processed based on the subject term and the preset weight of the subject term;
and the emotion analysis result determining module is used for determining the emotion analysis result of the text to be processed based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value.
Optionally, the second emotion intensity determination module is configured to:
dividing the whole sentence in the text to be processed into clauses according to punctuations in the text to be processed;
determining a fourth emotion intensity value of the clause;
a second sentimental intensity value is determined based on the fourth sentimental intensity value.
Optionally, when determining the fourth emotion intensity value of the clause, the second emotion intensity determining module is specifically configured to:
determining the emotional words, the negative words for modifying the emotional words and the degree adverbs for modifying the emotional words in the clauses;
determining sentence patterns of the clauses;
and determining a fourth emotion intensity value based on the preset weight corresponding to the emotion words, the preset weight coefficient corresponding to the negative words and the preset weight coefficient corresponding to the degree adverbs and the preset weight coefficient corresponding to the sentence patterns.
Optionally, when determining the second emotion intensity value based on the fourth emotion intensity value, the second emotion intensity determining module is specifically configured to:
determining the sentence relation between each clause and the adjacent clauses based on the associated words in each clause in the text to be processed;
and determining a second emotion intensity value based on the fourth emotion intensity value and a preset weight coefficient corresponding to the inter-sentence relation.
Optionally, the third emotion intensity determination module is specifically configured to:
determining a fifth emotion intensity value of the text to be processed based on the first emotion intensity value, a preset weight coefficient corresponding to the first emotion intensity value, the second emotion intensity value, a preset weight coefficient corresponding to the second emotion intensity value, the third emotion intensity value, a preset weight coefficient corresponding to the third emotion intensity value and a preset emotion intensity correction coefficient;
and determining an emotion analysis result of the text to be processed based on the fifth emotion intensity value.
Optionally, the emotion analysis result of the text to be processed includes an emotion polarity of the text to be processed, and the third emotion intensity determining module is specifically configured to, when determining the emotion analysis result of the text to be processed based on the fifth emotion intensity value:
and determining the emotion polarity of the text to be processed based on the fifth emotion intensity value and a preset emotion intensity threshold value.
In a third aspect, an embodiment of the present application provides an electronic device, including: a processor and a memory;
a memory for storing operating instructions;
a processor configured to execute the method according to any of the embodiments of the first aspect of the present application by calling an operation instruction.
In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the method shown in any implementation manner of the first aspect of the present application.
The technical scheme provided by the embodiment of the application has the following beneficial effects:
according to the scheme provided by the embodiment of the application, a first emotion intensity value of a to-be-processed text is determined based on a Bi-directional Long Short-Term Memory model, a second emotion intensity value of the to-be-processed text is determined based on an emotion expression rule, a third emotion intensity value of the to-be-processed text is determined based on a subject word and a subject word weight of the to-be-processed text, so that an emotion analysis result of the to-be-processed text is determined based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments of the present application will be briefly described below.
Fig. 1 is a schematic flowchart of a method for emotion analysis of a text according to an embodiment of the present application;
FIG. 2 is a schematic flow chart diagram of an embodiment of the present application;
FIG. 3 is a schematic diagram of a training process of a word2vector model in an embodiment of the present application;
FIG. 4 is a schematic diagram of a training process of a BilSTM model in an embodiment of the present application;
FIG. 5 is a schematic flow chart illustrating parameter adjustment and model testing of a BilSTM model according to an embodiment of the present application;
FIG. 6 is a schematic structural diagram of an emotion analysis system for a text to be processed according to an embodiment of the present application;
fig. 7 is a schematic structural diagram of an emotion analyzing apparatus for text according to an embodiment of the present application;
fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary only for the purpose of explaining the present application and are not to be construed as limiting the present invention.
As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. As used herein, the term "and/or" includes all or any element and all combinations of one or more of the associated listed items.
To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
The following describes the technical solutions of the present application and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.
Fig. 1 shows a schematic flow diagram of a text emotion analysis method provided in an embodiment of the present application, and as shown in fig. 1, the method mainly includes:
step S110: and determining a first emotion intensity value of the text to be processed based on a Bi-directional Long Short-Term Memory (BilSTM) model.
In the embodiment of the application, the BilSTM model can be trained through the labeled data set to obtain the trained BilSTM model. The corresponding word vector of the text to be processed can be obtained, the word vector of the text to be processed is input into the trained BilSTM model, and the first emotion intensity value of the text to be processed is determined according to the output result.
In the embodiment of the application, a word2vector model can be adopted to train a text to be processed to obtain a word vector corresponding to the text to be processed, the trained word vector contains semantic information of words in the text to be processed in mass data, and the information of the words in the text to be processed can be better represented through the word vector.
Step S120: and determining a second emotion intensity value of the text to be processed based on a predetermined emotion expression rule.
In the embodiment of the application, emotion expression rules can be preset to perform emotion analysis on the text to be processed, and specifically, the second emotion intensity value of the text to be processed can be determined by presetting an emotion word dictionary, semantic rules and the like.
Step S130: and acquiring a subject term of the text to be processed, and determining a third emotional intensity value of the text to be processed based on the subject term and a preset subject term weight.
In the embodiment of the application, the weight of the subject word can be preset through a preset subject word dictionary, the subject word of the text to be processed is extracted, and the weight of the extracted subject word is determined, so that the third emotional intensity value of the text to be processed is determined.
The subject term can represent the subject of the text to be processed, and the emotional intensity value of the text to be processed can be effectively determined based on the subject term and the weight of the subject term.
In the embodiment of the application, the topic word dictionary can be obtained by clustering massive positive short texts and negative short texts and then extracting the topic and the weight of each positive and negative category.
In the embodiment of the application, a preset number of subject words before a text can be extracted through an underlying Dirichlet Allocation (LDA) model, the extracted subject words are matched with a subject word dictionary to obtain corresponding weights, and a third emotional intensity value of the text to be processed is determined based on the weights of the extracted subject words.
Step S140: and determining the emotion analysis result of the text to be processed based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value.
In the embodiment of the application, the emotion analysis result of the text to be processed may be determined based on the determination of the first emotion intensity value, the second emotion intensity value and the third emotion intensity value, and the emotion analysis result may include the emotion polarity of the text to be processed.
The method provided by the embodiment of the application determines a first emotion intensity value of a to-be-processed text based on a BilSTM model, determines a second emotion intensity value of the to-be-processed text based on emotion expression rules, and determines a third emotion intensity value of the to-be-processed text based on subject words and subject word weights of the to-be-processed text, so that an emotion analysis result of the to-be-processed text is determined based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value.
In practical use, because the BilSTM model needs to be trained through a preset data set, the trained BilSTM model has higher accuracy when analyzing the text to be processed in the same field as the preset data set, and has lower accuracy when analyzing the text to be processed in the field different from the preset data set, but the data set used for large-scale training is rare at present, so if the emotion analysis is performed on the text to be processed only based on the BilSTM model, the accuracy and the qualification are not high enough when analyzing the text to be processed in the cross-field.
When emotion analysis is performed on a text to be processed based on a preset emotion expression rule, the emotion word dictionary and the semantic rule are established manually, so that the period for establishing the emotion word dictionary and the semantic rule is long, a large amount of manpower is consumed, and the established emotion word dictionary and the semantic rule are poor in transportability.
According to the method, the analysis modes based on the BilSTM model and the emotion expression rule are fused, the analysis modes based on the subject words and the subject word weights are fused, and the advantages of artificial experience knowledge and machine learning can be fully utilized to improve the accuracy of emotion analysis.
Fig. 2 is a flowchart illustrating a specific implementation manner of an embodiment of the present application, and in fig. 2, a text set is a text to be processed. And determining a first emotion intensity value of the text to be processed by a BilSTM model in the emotion prediction module. And determining a second emotion intensity value of the text to be processed based on the rule-based emotion dictionary in the emotion prediction module, namely based on the preset emotion expression rule. And determining a third emotion intensity value of the text to be processed based on the theme words and the preset theme word weight in the emotion prediction module. And (4) fusion, namely performing emotion analysis based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value. And the emotion prediction service is used for determining the emotion analysis result of the text to be processed. And determining the emotion polarity of the text to be processed based on the emotion analysis result.
In an embodiment of the present application, training the BilSTM model may include performing word vector training on a predetermined data set, performing preprocessing on the predetermined data set, training the BilSTM model, and performing word vector training on the predetermined data set,
(1) And performing word vector training on the preset data set.
In the embodiment of the application, word vector training is performed on a preset data set by adopting a word2vector model.
Usually, when model training is performed, Chinese in a data set is converted into a digital form to be accepted by a model, and Chinese of a text to be processed is converted into a data form, namely word embedding. The conventional word embedding is performed by word2id, i.e. dividing words in a data set, counting word frequency, and sorting according to word frequency, and finally obtaining a dictionary of all words. However, the word representation method of word2id only considers the word frequency and ignores important information among many words, so that the word representation method is insufficient or too simple, thereby also affecting the accuracy of emotion analysis. Therefore, a more complex pre-training model word2vector is introduced, which is a method of word representation based on the context of words. The training flow of the word2vector model is shown in FIG. 3. In fig. 3, a corpus is loaded, that is, a data set for word2vector model training is loaded. And (4) jieba word segmentation, namely carrying out word segmentation on the loaded data set through a jieba component. word2vector model, the data set after word segmentation is input into the word2vector model. And setting parameters, namely setting parameters of the word2vector model in the training process. And (4) saving the model, namely saving the word2vector model after parameter setting is finished.
(2) Preprocessing a preset data set
Interference information influencing semantic continuity may exist in the preset data set, such as characters containing link parts, coded special characters, semantic-free information and the like. Thus, preprocessing the preset data set can filter out such interference information.
(3) Training of BilSTM models
①, loading a preset data set for emotion polarity labeling, performing jieba word segmentation on the preset data set, loading a trained word2vector model, and converting words into word vectors based on the word2vector model.
② since the input of the recurrent neural network is a fixed time sequence length, that is, the data length of each input is fixed, the fixed data length can be preset, since there may be more short texts not meeting the preset data length in the data set, we with the length less than the preset data length will fill the preset data length with 0 in order to keep the length of all input sequences consistent.
In actual use, the input sequence may be marked as an X list, 1 is set to indicate positive emotion polarity, 0 is set to indicate negative emotion polarity, and the emotion label corresponding to the word vector is marked as a Y list. Taking an X list and a Y list as input.
③ the BiLSTM model is constructed by four layers of network structure, the first layer is the input layer, the second layer is the BiLSTM layer which mainly uses two-way memory to learn the probability distribution of the context of the text, the third layer is the regular layer which adds Dropout operation to train the simplified network to prevent overfitting, the fourth layer is the full connection layer which adds a linear operation to activate the network (sigmoid) for function, which is used to convert the matrix into 2-dimensional output, i.e. 2 classification, by the matrix operation of the full connection layer, then the probability distribution of the final two classifications is obtained by the sigmoid.
④, updating model parameters for constructing the BilSTM model, calculating the error between the model class output and the real class, namely loss, through cross entropy, then optimizing the loss part through an adam optimizer, searching the point with the minimum loss through a random gradient descent method, updating the model parameters through derivation, and circulating for many times until the data training is completed.
⑤, storing the parameters of the training update and storing the constructed BilSTM model.
The training process of the BilSTM model is shown in FIG. 4, and as shown in FIG. 4, the data set is labeled, that is, the emotion polarity of the preset data set is labeled. And loading the word vector model, namely loading the trained word2vector model. Word embedding (wordemmbedding), i.e. after a preset data set is subjected to jieba word segmentation, the determined words are converted into word vectors. Padding (padding), i.e. padding the length of the input word vector so that it meets the preset data length. And the network layer inputs the filled data into a BilSTM model. Binary cross entropy (binary _ cross), i.e. the error of the model class output and the real class is calculated by cross entropy. Adam, i.e., the loss fraction is optimized by an Adam optimizer. Accuracy (accuracy), i.e. the accuracy with which the output result is determined. The number of trainings (epochs), i.e., the number of trainings for which the BilSTM model is trained. The Save model (Save model), i.e., the BilSTM model at the end of training, is saved. Acc > e epochs > f, namely when the accuracy is greater than a preset accuracy threshold e or the training times are greater than a time threshold f, the training is finished. And (4) the epochs are less than f or Acc < e, namely when the training times are less than the time threshold f and the accuracy is less than the preset accuracy threshold e, the training is carried out again.
(4) Model testing is carried out on the trained BilSTM model
Model testing was performed on an entirely new dataset, different from the dataset used during training, to ensure that the generalization capability of the BilSTM model was tested. The performance parameter selected during testing may be accuracy (accuracy). The accuracy rate represents the proportion of the predicted emotion polarity of the text to be processed and the actual emotion polarity of the text to be processed, which are consistent.
Parameter adjustment and model test of the BilSTM model can be performed together, and the specific flow can be as shown in FIG. 5. In fig. 5, initialization, namely, initialization is performed on the constructed BiLSTM model. And (4) model training, namely training the constructed BilSTM model. And (4) satisfying the training stopping condition, namely outputting a result satisfying the training stopping condition when the constructed BilSTM model is trained. And (4) selecting hyper-parameters, namely adjusting model parameters and updating the model parameters in the training process. And (4) model testing, namely performing model testing on the trained BilSTM model after model training is completed.
In an optional manner of the embodiment of the present application, determining a second emotion intensity value of a text to be processed based on a predetermined emotion expression rule includes:
dividing the whole sentence in the text to be processed into clauses according to punctuations in the text to be processed;
determining a fourth emotion intensity value of the clause;
a second sentimental intensity value is determined based on the fourth sentimental intensity value.
In the embodiment of the application, the text to be processed may include a plurality of whole sentences, and each whole sentence may be divided into clauses based on punctuation marks in the text to be processed. For example, clauses may be divided by commas in the text to be processed, i.e., sentences separated by commas are determined as clauses.
Because each clause in the text to be processed may be capable of performing emotional expression, a fourth emotional intensity value of each clause may be determined, and then a second emotional intensity value of the text to be processed is determined based on the fourth emotional intensity value of each sentence.
In an optional manner of the embodiment of the present application, determining the fourth emotion intensity value of the clause includes:
determining the emotional words, the negative words for modifying the emotional words and the degree adverbs for modifying the emotional words in the clauses;
determining sentence patterns of the clauses;
and determining a fourth emotion intensity value based on the preset weight corresponding to the emotion words, the preset weight coefficient corresponding to the negative words and the preset weight coefficient corresponding to the degree adverbs and the preset weight coefficient corresponding to the sentence patterns.
In the embodiment of the application, the emotion words in the clauses are words directly expressing emotion, the emotion intensities expressed by different emotion words are different, and the weights can be preset for the emotion words in the emotion word dictionary respectively.
When the emotional words are modified by the degree adverbs, the degree adverbs can strengthen or weaken the emotional intensity of the clauses, and weight coefficients can be preset for the degree adverbs in the emotional word dictionary.
When the emotion words are modified by the negative words, the negative words mean that the polarities of the emotions of the clauses are changed, so that the weight coefficient of the negative words can be preset in the emotion word dictionary, and the weight coefficient of the negative words is negative one.
In the embodiment of the present application, the sentence pattern may include: statement sentences, exclamation sentences, question reversals, and the like. Clause patterns can also affect the emotional tendency of clauses. For example, when a clause is a question-back sentence, the effect on the expressed emotion of the clause is a reverse reinforcement. For example, a clause is an exclamation, and the effect on the expressed emotion of the clause is to enhance the emotional tendency of the clause. When calculating the emotion intensity of a clause, different weight coefficients may be set for each sentence pattern. For example, the emotive intensity coefficient of the exclamation sentence may be 2, and the emotive intensity coefficient of the question sentence may be-2.
In actual use, the sentence pattern of the clause may be determined by the end punctuation of the clause where the clause ends, e.g., the clause ends in "! "in time, the sentence pattern is an exclamation sentence. The sentence pattern of the clause can also be determined as a question reversing sentence by a reverse question word in the clause, such as "difficult sentence".
In actual use, the fourth emotion intensity value may be determined based on the following formula:
in the above formula (1), H (w)i) Expressing a fourth emotion intensity value, i expressing any emotion word in the clause, n expressing the total number of emotion words in the clause, wiWeight of any one of the above emotional words, neg weight coefficient of a negative word that modifies the above emotional word in a clause, d weight coefficient of an adverb that modifies the above emotional word in a clause, MrAnd weight coefficients corresponding to the sentence patterns of the clauses are shown.
In an optional manner of the embodiment of the present application, determining the second emotion intensity value based on the fourth emotion intensity value includes:
determining the sentence relation between each clause and the adjacent clauses based on the associated words in each clause in the text to be processed;
and determining a second emotion intensity value based on the fourth emotion intensity value and a preset weight coefficient corresponding to the inter-sentence relation.
In the embodiment of the application, the text to be processed may include a plurality of clauses, and adjacent clauses may be related to each other, that is, inter-sentence relationships may exist. The inter-sentence relationship may be a turning relationship, a hypothetical relationship, a causal relationship, or the like. In actual use, the inter-sentence relationship can be determined by the related words in the clauses, for example, the turning connection words "although" and "but" between the front and back clauses are used to determine the turning relationship between the front and back clauses.
Since the inter-sentence relationship may affect the emotion expressed by the preceding and following clauses, a weight coefficient may be preset for the inter-sentence relationship. Since the inter-sentence relationship may have different influences on the emotion expressed by the preceding and following clauses, the preset weight coefficient in the inter-sentence relationship may be different for each clause constituting the inter-sentence relationship, for example, the inter-sentence relationship is a turning relationship, and the weight coefficients of the clause before turning and the clause after turning may be different.
In the embodiment of the present application, because the influence of the inter-sentence relation is taken into consideration, a plurality of clauses having inter-sentence relation may be determined as one clause group, and the sixth emotion intensity value of the clause group may be determined by calculating the fourth emotion intensity value of each clause in the sentence group. And determining a second emotion intensity value based on the sixth emotion intensity value of the clause group and the fourth emotion intensity value of the clause which has no inter-sentence relation with other clauses.
In actual use, a second emotional intensity value is determined:
in the above formula (2), SdExpressing a second emotion intensity value, k expressing any clause group with an inter-sentence relation in the text to be processed, m expressing the total number of the clause groups in the text to be processed, l expressing any clause in any clause group, u expressing the total number of the clauses in any clause group, Hku(wi) A fourth emotion intensity value f representing any clause in any of the clause groupskRepresenting the weight coefficient corresponding to any clause in any clause group,
the sixth emotion intensity values of all clause groups in the text to be processed are represented, s represents any clause which has no inter-sentence relation with other clauses in the text to be processed, r represents the total number of clauses which have no inter-sentence relation with other clauses in the text to be processed, and H represents the total number of clauses which have no inter-sentence relation with other clauses in the text to be processeds(wi) And a fourth emotion intensity value indicating any clause which does not have an inter-sentence relationship with other clauses.
In practical use of the embodiment of the application, determining the third emotion intensity value of the text to be processed based on the subject term and the preset subject term weight may include:
determining a third sentiment intensity value based on the following formula:
in the above formula (3), StRepresenting a third emotional intensity value, c representing any subject word obtained from the text to be processed, v the total number of subject words obtained from the text to be processed, tcRepresenting the weight of any of the subject words described above.
In an optional manner of the embodiment of the present application, determining an emotion analysis result of a to-be-processed text based on a first emotion intensity value, a second emotion intensity value, and a third emotion intensity value includes:
determining a fifth emotion intensity value of the text to be processed based on the first emotion intensity value, a preset weight coefficient corresponding to the first emotion intensity value, the second emotion intensity value, a preset weight coefficient corresponding to the second emotion intensity value, the third emotion intensity value, a preset weight coefficient corresponding to the third emotion intensity value and a preset emotion intensity correction coefficient;
and determining an emotion analysis result of the text to be processed based on the fifth emotion intensity value.
In the embodiment of the application, weighting coefficients can be respectively preset for the first emotion intensity value, the second emotion intensity value and the third emotion intensity value, so that the first emotion intensity value, the second emotion intensity value and the third emotion intensity value are subjected to weighted calculation, and a fifth emotion intensity value is determined.
When the first emotion intensity value is determined through the BilSTM model, the requirement on the data set for training is high, and when the coverage degree of the data set for training on the field to which the text to be processed belongs is low, the accuracy of the determined first emotion intensity value intensity is poor. When the second emotion intensity value is determined through the emotion expression rule, the preset emotion expression rule can cover all fields, but the preset emotion expression rule has no learning function and is low in accuracy. Therefore, when the coverage degree of the training data set to the field of the text to be processed is low, a low weight coefficient can be preset for the first emotion intensity value; and when the coverage degree of the data set used for training to the prediction field is higher and the accuracy of the determined first emotion intensity value intensity is higher, a higher weight coefficient can be preset for the first emotion intensity value.
In actual use, the fifth emotional intensity value may be determined based on the following formula:
S＝λ1SR+λ2St+λ3Sd+b (4)
in the above formula (4), S represents a fifth emotion intensity value, λ1Weight coefficient representing first emotion intensity value, SRRepresenting a first emotional intensity value, λ2Weight coefficient representing second emotional intensity value, StRepresenting a second emotional intensity value, λ3Weight coefficient representing third emotion intensity value, SdThe weight coefficient of the third emotion value, b represents a correction coefficient of emotion intensity.
In an optional manner of the embodiment of the present application, the emotion analysis result of the to-be-processed text includes an emotion polarity of the to-be-processed text, and the determining the emotion analysis result of the to-be-processed text based on the fifth emotion intensity value includes:
and determining the emotion polarity of the text to be processed based on the fifth emotion intensity value and a preset emotion intensity threshold value.
In the embodiment of the application, the emotion analysis result of the text to be processed may be the emotion polarity of the text to be processed, specifically, an emotion intensity threshold may be preset, and when a fifth emotion intensity value of the text to be processed is greater than the preset emotion intensity threshold, the emotion polarity of the text to be processed may be determined to be a forward direction; when the fifth emotion intensity value of the text to be processed is not greater than the preset emotion intensity threshold value, it may be determined that the emotion polarity of the text to be processed is negative.
Fig. 6 shows a schematic structural diagram of an emotion analysis system for a text to be processed, which adopts a distributed Spark cluster to implement emotion analysis for the text to be processed in consideration of throughput requirements and actual text data volume. The emotion analysis System of the text to be processed is managed through a zookeeper service, the text to be processed is written into a Hadoop Distributed File System (HDFS) in real time through an input module, a scattered data stream is read from the HDFS through a Spark Streaming, the emotion analysis method of the text is executed, emotion analysis is carried out on the text to be processed, emotion analysis results are written into a message queue, specifically, the emotion analysis results can be returned to a kafka topic in the form of the scattered data stream, and kafka data is read through an application program to be displayed on a webpage.
Based on the same principle as the method shown in fig. 1, fig. 7 shows a schematic structural diagram of an emotion analysis apparatus for text according to an embodiment of the present application, and as shown in fig. 7, the emotion analysis apparatus 20 for text may include:
the first emotion intensity determining module 210 is configured to determine a first emotion intensity value of the text to be processed based on the BiLSTM model;
the second emotion intensity determination module 220 is configured to determine a second emotion intensity value of the text to be processed based on a predetermined emotion expression rule;
a third emotion intensity determining module 230, configured to obtain a subject term of the to-be-processed text, and determine a third emotion intensity value of the to-be-processed text based on the subject term and a preset subject term weight;
and an emotion analysis result determining module 240, configured to determine an emotion analysis result of the to-be-processed text based on the first emotion intensity value, the second emotion intensity value, and the third emotion intensity value.
The device provided by the embodiment of the application determines a first emotion intensity value of a to-be-processed text based on a BilSTM model, determines a second emotion intensity value of the to-be-processed text based on emotion expression rules, and determines a third emotion intensity value of the to-be-processed text based on subject words and subject word weights of the to-be-processed text, so that emotion analysis results of the to-be-processed text are determined based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value.
Optionally, the second emotion intensity determination module is configured to:
dividing the whole sentence in the text to be processed into clauses according to punctuations in the text to be processed;
determining a fourth emotion intensity value of the clause;
a second sentimental intensity value is determined based on the fourth sentimental intensity value.
Optionally, when determining the fourth emotion intensity value of the clause, the second emotion intensity determining module is specifically configured to:
determining the emotional words, the negative words for modifying the emotional words and the degree adverbs for modifying the emotional words in the clauses;
determining sentence patterns of the clauses;
and determining a fourth emotion intensity value based on the preset weight corresponding to the emotion words, the preset weight coefficient corresponding to the negative words and the preset weight coefficient corresponding to the degree adverbs and the preset weight coefficient corresponding to the sentence patterns.
Optionally, when determining the second emotion intensity value based on the fourth emotion intensity value, the second emotion intensity determining module is specifically configured to:
determining the sentence relation between each clause and the adjacent clauses based on the associated words in each clause in the text to be processed;
and determining a second emotion intensity value based on the fourth emotion intensity value and a preset weight coefficient corresponding to the inter-sentence relation.
Optionally, the third emotion intensity determination module is specifically configured to:
determining a fifth emotion intensity value of the text to be processed based on the first emotion intensity value, a preset weight coefficient corresponding to the first emotion intensity value, the second emotion intensity value, a preset weight coefficient corresponding to the second emotion intensity value, the third emotion intensity value, a preset weight coefficient corresponding to the third emotion intensity value and a preset emotion intensity correction coefficient;
and determining an emotion analysis result of the text to be processed based on the fifth emotion intensity value.
Optionally, the emotion analysis result of the text to be processed includes an emotion polarity of the text to be processed, and the third emotion intensity determining module is specifically configured to, when determining the emotion analysis result of the text to be processed based on the fifth emotion intensity value:
and determining the emotion polarity of the text to be processed based on the fifth emotion intensity value and a preset emotion intensity threshold value.
It can be understood that the above modules of the text emotion analysis apparatus in the embodiment have functions of implementing the corresponding steps of the text emotion analysis method in the embodiment shown in fig. 1. The function can be realized by hardware, and can also be realized by executing corresponding software by hardware. The hardware or software includes one or more modules corresponding to the functions described above. The modules can be software and/or hardware, and each module can be implemented independently or by integrating a plurality of modules. For the functional description of each module of the text emotion analysis apparatus, reference may be specifically made to the corresponding description of the text emotion analysis method in the embodiment shown in fig. 1, and details are not repeated here.
The embodiment of the application provides an electronic device, which comprises a processor and a memory;
a memory for storing operating instructions;
and the processor is used for executing the emotion analysis method of the text provided by any embodiment of the application by calling the operation instruction.
As an example, fig. 8 shows a schematic structural diagram of an electronic device to which an embodiment of the present application is applicable, and as shown in fig. 8, the electronic device 2000 includes: a processor 2001 and a memory 2003. Wherein the processor 2001 is coupled to a memory 2003, such as via a bus 2002. Optionally, the electronic device 2000 may also include a transceiver 2004. It should be noted that the transceiver 2004 is not limited to one in practical applications, and the structure of the electronic device 2000 is not limited to the embodiment of the present application.
The processor 2001 is applied to the embodiment of the present application to implement the method shown in the above method embodiment. The transceiver 2004 may include a receiver and a transmitter, and the transceiver 2004 is applied to the embodiments of the present application to implement the functions of the electronic device of the embodiments of the present application to communicate with other devices when executed.
The Processor 2001 may be a CPU (Central Processing Unit), general Processor, DSP (Digital Signal Processor), ASIC (Application specific integrated Circuit), FPGA (Field Programmable Gate Array) or other Programmable logic device, transistor logic device, hardware component, or any combination thereof. Which may implement or perform the various illustrative logical blocks, modules, and circuits described in connection with the disclosure. The processor 2001 may also be a combination of computing functions, e.g., comprising one or more microprocessors, DSPs and microprocessors, and the like.
The Memory 2003 may be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, an EEPROM (Electrically erasable programmable Read Only Memory), a CD-ROM (Compact disk Read Only Memory) or other optical disk storage, optical disk storage (including Compact disk, laser disk, optical disk, digital versatile disk, blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to these.
Optionally, the memory 2003 is used for storing application program code for performing the disclosed aspects, and is controlled in execution by the processor 2001. The processor 2001 is configured to execute the application program code stored in the memory 2003 to implement the emotion analysis method for text provided in any of the embodiments of the present application.
The electronic device provided by the embodiment of the application is applicable to any embodiment of the method, and is not described herein again.
Compared with the prior art, the electronic equipment has the advantages that the first emotion intensity value of the text to be processed is determined based on the BilSTM model, the second emotion intensity value of the text to be processed is determined based on the emotion expression rule, the third emotion intensity value of the text to be processed is determined based on the subject word and the subject word weight of the text to be processed, the emotion analysis result of the text to be processed is determined based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value, the emotion analysis result of the text to be processed is rapidly obtained through analysis of the text to be processed, and the emotion polarity of netizen release information can be timely obtained.
The embodiment of the application provides a computer readable storage medium, and a computer program is stored on the computer readable storage medium, and when the program is executed by a processor, the computer readable storage medium implements the emotion analysis method for a text shown in the above method embodiment.
The computer-readable storage medium provided in the embodiments of the present application is applicable to any of the embodiments of the foregoing method, and is not described herein again.
Compared with the prior art, the embodiment of the application provides a computer-readable storage medium, a first emotion intensity value of a to-be-processed text is determined based on a BilSTM model, a second emotion intensity value of the to-be-processed text is determined based on emotion expression rules, a third emotion intensity value of the to-be-processed text is determined based on a subject word and a subject word weight of the to-be-processed text, so that an emotion analysis result of the to-be-processed text is determined based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The foregoing is only a partial embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A method for emotion analysis of a text, comprising:
determining a first emotion intensity value of a text to be processed based on a Bi-directional Long Short-Term Memory model;
determining a second emotion intensity value of the text to be processed based on a preset emotion expression rule;
acquiring a subject term of the text to be processed, and determining a third emotion intensity value of the text to be processed based on the subject term and a preset subject term weight;
and determining the emotion analysis result of the text to be processed based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value.
2. The method of claim 1, wherein the determining a second emotion intensity value of the text to be processed based on a predetermined emotion expression rule comprises:
dividing the whole sentence in the text to be processed into clauses according to punctuations in the text to be processed;
determining a fourth emotion intensity value of the clause;
determining the second sentiment intensity value based on the fourth sentiment intensity value.
3. The method of claim 2, wherein said determining a fourth emotion intensity value for said clause comprises:
determining emotional words, negative words for modifying the emotional words and degree adverbs for modifying the emotional words in the clauses;
determining sentence patterns of the clauses;
and determining the fourth emotion intensity value based on the preset weight corresponding to the emotion words, the preset weight coefficient corresponding to the negative words and the preset weight coefficient corresponding to the degree adverbs and the preset weight coefficient corresponding to the sentence patterns.
4. The method of claim 2, wherein determining the second sentiment intensity value based on the fourth sentiment intensity value comprises:
determining the inter-sentence relation between each clause and the adjacent clause based on the associated words in each clause in the text to be processed;
and determining the second emotion intensity value based on the fourth emotion intensity value and a preset weight coefficient corresponding to the sentence-to-sentence relation.
5. The method of claim 1, wherein determining the emotion analysis result of the text to be processed based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value comprises:
determining a fifth emotion intensity value of the text to be processed based on the first emotion intensity value, a preset weight coefficient corresponding to the first emotion intensity value, a second emotion intensity value, a preset weight coefficient corresponding to the second emotion intensity value, a third emotion intensity value, a preset weight coefficient corresponding to the third emotion intensity value and a preset emotion intensity correction coefficient;
and determining an emotion analysis result of the text to be processed based on the fifth emotion intensity value.
6. The method of claim 5, wherein the emotion analysis result of the text to be processed comprises an emotion polarity of the text to be processed, and wherein the determining the emotion analysis result of the text to be processed based on the fifth emotion intensity value comprises:
and determining the emotion polarity of the text to be processed based on the fifth emotion intensity value and a preset emotion intensity threshold value.
7. An emotion analysis device for a text, comprising:
the first emotional intensity determining module is used for determining a first emotional intensity value of the text to be processed based on the Bi-directional Long Short-Term Memory model;
the second emotion intensity determination module is used for determining a second emotion intensity value of the text to be processed based on a preset emotion expression rule;
the third emotional intensity determination module is used for acquiring the subject term of the text to be processed and determining a third emotional intensity value of the text to be processed based on the subject term and a preset subject term weight;
and the emotion analysis result determining module is used for determining the emotion analysis result of the text to be processed based on the first emotion intensity value, the second emotion intensity value and the third emotion intensity value.
8. The apparatus of claim 7, wherein the second emotion intensity determination module is configured to:
dividing the whole sentence in the text to be processed into clauses according to punctuations in the text to be processed;
determining a fourth emotion intensity value of the clause;
determining the second sentiment intensity value based on the fourth sentiment intensity value.
9. An electronic device comprising a processor and a memory;
the memory is used for storing operation instructions;
the processor is used for executing the method of any one of claims 1-6 by calling the operation instruction.
10. A computer-readable storage medium, characterized in that the storage medium has stored thereon a computer program which, when being executed by a processor, carries out the method of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911156648.7A CN110929516A (en) | 2019-11-22 | 2019-11-22 | Text emotion analysis method and device, electronic equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911156648.7A CN110929516A (en) | 2019-11-22 | 2019-11-22 | Text emotion analysis method and device, electronic equipment and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110929516A true CN110929516A (en) | 2020-03-27 |
Family
ID=69850739
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911156648.7A Pending CN110929516A (en) | 2019-11-22 | 2019-11-22 | Text emotion analysis method and device, electronic equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110929516A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112507337A (en) * | 2020-12-18 | 2021-03-16 | 四川长虹电器股份有限公司 | Implementation method of malicious JavaScript code detection model based on semantic analysis |
| CN112733525A (en) * | 2021-01-18 | 2021-04-30 | 北京奇艺世纪科技有限公司 | Object emotion analysis method and device |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103744953A (en) * | 2014-01-02 | 2014-04-23 | 中国科学院计算机网络信息中心 | Network hotspot mining method based on Chinese text emotion recognition |
| US20150286627A1 (en) * | 2014-04-03 | 2015-10-08 | Adobe Systems Incorporated | Contextual sentiment text analysis |
| CN105095190A (en) * | 2015-08-25 | 2015-11-25 | 众联数据技术（南京）有限公司 | Chinese semantic structure and finely segmented word bank combination based emotional analysis method |
| CN107862087A (en) * | 2017-12-01 | 2018-03-30 | 广州简亦迅信息科技有限公司 | Sentiment analysis method, apparatus and storage medium based on big data and deep learning |
| CN107908782A (en) * | 2017-12-06 | 2018-04-13 | 陕西识代运筹信息科技股份有限公司 | A kind of data processing method and device based on sentiment analysis |
| CN108268439A (en) * | 2016-12-30 | 2018-07-10 | 北京国双科技有限公司 | The processing method and processing device of text emotion |
| CN108536681A (en) * | 2018-04-16 | 2018-09-14 | 腾讯科技（深圳）有限公司 | Intelligent answer method, apparatus, equipment and storage medium based on sentiment analysis |
| 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 |
-
2019
- 2019-11-22 CN CN201911156648.7A patent/CN110929516A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103744953A (en) * | 2014-01-02 | 2014-04-23 | 中国科学院计算机网络信息中心 | Network hotspot mining method based on Chinese text emotion recognition |
| US20150286627A1 (en) * | 2014-04-03 | 2015-10-08 | Adobe Systems Incorporated | Contextual sentiment text analysis |
| CN105095190A (en) * | 2015-08-25 | 2015-11-25 | 众联数据技术（南京）有限公司 | Chinese semantic structure and finely segmented word bank combination based emotional analysis method |
| CN108268439A (en) * | 2016-12-30 | 2018-07-10 | 北京国双科技有限公司 | The processing method and processing device of text emotion |
| CN107862087A (en) * | 2017-12-01 | 2018-03-30 | 广州简亦迅信息科技有限公司 | Sentiment analysis method, apparatus and storage medium based on big data and deep learning |
| CN107908782A (en) * | 2017-12-06 | 2018-04-13 | 陕西识代运筹信息科技股份有限公司 | A kind of data processing method and device based on sentiment analysis |
| CN108536681A (en) * | 2018-04-16 | 2018-09-14 | 腾讯科技（深圳）有限公司 | Intelligent answer method, apparatus, equipment and storage medium based on sentiment analysis |
| 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 |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112507337A (en) * | 2020-12-18 | 2021-03-16 | 四川长虹电器股份有限公司 | Implementation method of malicious JavaScript code detection model based on semantic analysis |
| CN112733525A (en) * | 2021-01-18 | 2021-04-30 | 北京奇艺世纪科技有限公司 | Object emotion analysis method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110609899B (en) | Specific target emotion classification method based on improved BERT model | |
| CN111985229B (en) | Sequence labeling method and device and computer equipment | |
| CN110781686B (en) | Statement similarity calculation method and device and computer equipment | |
| CN113987187B (en) | Public opinion text classification method, system, terminal and medium based on multi-label embedding | |
| CN110442721B (en) | Neural network language model, training method, device and storage medium | |
| WO2023137911A1 (en) | Intention classification method and apparatus based on small-sample corpus, and computer device | |
| CN111522908A (en) | Multi-label text classification method based on BiGRU and attention mechanism | |
| CN108416032A (en) | A kind of file classification method, device and storage medium | |
| CN109726400B (en) | Entity word recognition result evaluation method, device, equipment and entity word extraction system | |
| CN115310443A (en) | Model training method, information classification method, device, equipment and storage medium | |
| US20220366145A1 (en) | Generative language model for few-shot aspect-based sentiment analysis | |
| CN111241232A (en) | Business service processing method and device, service platform and storage medium | |
| WO2023159756A1 (en) | Price data processing method and apparatus, electronic device, and storage medium | |
| CN110929516A (en) | Text emotion analysis method and device, electronic equipment and readable storage medium | |
| CN112287656A (en) | Text comparison method, device, equipment and storage medium | |
| CN114722832A (en) | Abstract extraction method, device, equipment and storage medium | |
| CN111723186A (en) | Knowledge graph generation method based on artificial intelligence for dialog system and electronic equipment | |
| CN111259673A (en) | Feedback sequence multi-task learning-based law decision prediction method and system | |
| CN116578671A (en) | Emotion-reason pair extraction method and device | |
| Bai et al. | Gated character-aware convolutional neural network for effective automated essay scoring | |
| CN116092101A (en) | Training method, image recognition method apparatus, device, and readable storage medium | |
| CN113761935B (en) | Short text semantic similarity measurement method, system and device | |
| CN109446518B (en) | Decoding method and decoder for language model | |
| CN113761874A (en) | Event reality prediction method and device, electronic equipment and storage medium | |
| Chauhan | Sentiment analysis of customer reviews using deep learning techniques |
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: 20200327 |