CN111506793A - Method for realizing energy power public opinion analysis processing based on emotion mining - Google Patents
Method for realizing energy power public opinion analysis processing based on emotion mining Download PDFInfo
- Publication number
- CN111506793A CN111506793A CN202010300741.7A CN202010300741A CN111506793A CN 111506793 A CN111506793 A CN 111506793A CN 202010300741 A CN202010300741 A CN 202010300741A CN 111506793 A CN111506793 A CN 111506793A
- Authority
- CN
- China
- Prior art keywords
- energy
- public opinion
- emotion
- mining
- power
- 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
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
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
-
- 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
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic 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
- G06N3/048—Activation functions
-
- 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
Abstract
The invention relates to a method for realizing energy and power public opinion analysis and processing based on emotion mining, which comprises the steps of collecting samples, collecting information through a web crawler, preprocessing texts, segmenting the texts through a segmentation toolkit, vectorizing the texts through word2vec technology, performing convolution operation, performing pooling operation, performing full-link operation, classifying energy themes and identifying emotional tendencies according to a convolutional neural network, and predicting the energy and power public opinion tendencies through L STM sequence mining network.
Description
Technical Field
The invention relates to the field of deep learning, in particular to the field of public opinion analysis, and specifically relates to a method for realizing energy and power public opinion analysis processing based on emotion mining.
Background
At present, various technologies are used for comment public opinion analysis at home and abroad, and the most important of the technologies is a content analysis technology and an emotion analysis technology. The content analysis technology is mainly classified into a keyword matching algorithm, a knowledge engineering method, and a statistical learning method. The emotion analysis technology is mainly divided into an emotion dictionary, machine learning and deep learning. Although the emotion dictionary is simple, the emotion dictionary cannot adapt to rapid change of internet content, the traditional machine learning method is biased to static content analysis, and the generalization migration capability is lacked on content analysis of dynamic change. In addition, there are not many relevant studies or implementations of public opinion tendency prediction for energy power development.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides the method for realizing the analysis and the processing of the energy power public sentiment based on the emotion mining, which is simple and convenient to operate, high in accuracy, strong in generalization capability and wide in application range.
In order to achieve the purpose, the method for realizing the energy power public opinion analysis processing based on emotion mining comprises the following steps:
the method for realizing the energy power public opinion analysis processing based on emotion mining is mainly characterized by comprising the following steps of:
(1) collecting samples, and collecting information through a web crawler;
(2) performing text preprocessing, and segmenting words of the text through a word segmentation toolkit;
(3) performing text vectorization through word2vec technology;
(4) performing convolution operation;
(5) performing pooling operation;
(6) performing full-link operation, and performing energy theme classification and emotional tendency identification according to the convolutional neural network;
(7) and (4) carrying out energy public opinion tendency prediction through an L STM sequence mining network.
Preferably, the step (2) specifically comprises the following steps:
and performing word segmentation on the energy text through a word segmentation toolkit, and performing word deletion and stop operation on the text.
Preferably, the step (3) specifically includes the following steps:
defining a word sequence, and splicing the trained word vectors to obtain a word vector matrix which is used as the input of the convolutional layer.
Preferably, the step (5) specifically comprises the following steps:
and performing maximum pooling operation on the feature map extracted by each filter, and obtaining a vector through one-layer convolution and one-layer pooling.
Preferably, the step (6) specifically includes the following steps:
(6.1) feeding the vectors into a full connection layer to obtain final feature extraction vectors;
(6.2) judging whether the energy comment content is classified or the energy comment emotion is classified, and if the energy comment content is classified, taking a Softmax function as an activation function to classify the energy comment content into thermal power, hydroelectric power, wind power, tidal power, solar energy, nuclear energy and mixed energy; if the energy comment emotion is classified, the Sigmoid function is used as an activation function and is classified into negative, neutral and positive.
Preferably, the step (7) specifically comprises the following steps:
(7.1) training L STM predictor model;
(7.2) prediction was performed by L STM predictor model.
By adopting the method for realizing the energy power public opinion analysis processing based on emotion mining, an analysis model aiming at the energy development trend of an energy report is realized, the model well evaluates the public opinion trend of a power source, has stronger topic analysis and sequence mining prediction capabilities, and can be expanded to other time sequence public opinion analysis and prediction fields. In addition, fine-grained optimization can be continuously carried out on the theme mode of the algorithm, multi-theme multi-sequence emotion comparison analysis is realized by analyzing the internal structure of the information news, more comprehensive public sentiment research on the power trend is developed, the research on the energy development trend can be mastered in time, the mutual evidence is obtained with the energy development quantitative analysis, the future energy development trend and law can be grasped more objectively and accurately, and theoretical guidance is provided for constructing an energy system.
Drawings
Fig. 1 is a schematic energy trend analysis flow diagram of the method for implementing energy public power opinion analysis based on emotion mining according to the present invention.
Fig. 2 is a flow diagram of a chinese text classification algorithm based on CNN of the method for implementing energy public opinion analysis based on emotion mining according to the present invention.
Fig. 3 is a block diagram of L STM cycle of the method for implementing the energy power public sentiment analysis based on emotion mining according to the present invention.
Detailed Description
In order to more clearly describe the technical contents of the present invention, the following further description is given in conjunction with specific embodiments.
The invention discloses a method for realizing energy power public opinion analysis processing based on emotion mining, which comprises the following steps of:
(1) collecting samples, and collecting information through a web crawler;
(2) performing text preprocessing, and segmenting words of the text through a word segmentation toolkit;
performing word segmentation on the energy text through a word segmentation toolkit, and performing word deletion and stop operation on the text;
(3) performing text vectorization through word2vec technology;
defining a word sequence, splicing the trained word vectors to obtain a word vector matrix which is used as the input of the convolutional layer;
(4) performing convolution operation;
(5) performing pooling operation;
performing maximum pooling operation on the feature map extracted by each filter, and obtaining a vector through one-layer convolution and one-layer pooling;
(6) performing full-link operation, and performing energy theme classification and emotional tendency identification according to the convolutional neural network;
(6.1) feeding the vectors into a full connection layer to obtain final feature extraction vectors;
(6.2) judging whether the energy comment content is classified or the energy comment emotion is classified, and if the energy comment content is classified, taking a Softmax function as an activation function to classify the energy comment content into thermal power, hydroelectric power, wind power, tidal power, solar energy, nuclear energy and mixed energy; if the energy comment emotions are classified, taking the Sigmoid function as an activation function, and classifying the Sigmoid function into negative, neutral and positive;
(7) energy public opinion tendency prediction is carried out through an L STM sequence mining network;
(7.1) training L STM predictor model;
(7.2) prediction was performed by L STM predictor model.
In the specific implementation mode of the invention, the text classification algorithm based on the CNN is provided for the problem of energy and power news information classification. The framework comprises the steps of text preprocessing, text vectorization, convolution operation, pooling operation, full-connection operation and classification, and can accurately classify the energy theme and the comment content emotion.
The invention also provides a trend prediction algorithm based on L STM aiming at the problem of global energy power public opinion prediction based on L STM trend prediction algorithm, wherein the algorithm designs a L STM cycle unit, defines an effective loss function in a training stage and can accurately predict the global energy power public opinion.
The method realizes word segmentation through a word segmentation toolkit, such as Jieba, N L TK and the like, realizes text vectorization through word2vec technology, realizes energy theme classification and emotional tendency recognition through a convolutional neural network, and realizes public sentiment tendency prediction of future energy power development through a L STM sequence mining network.
The invention discloses a method for realizing energy power public opinion analysis processing based on emotion mining, which comprises the following steps of:
(1) collecting samples, namely collecting energy and electric power news information by using a web crawler technology;
(2) text preprocessing, namely performing word segmentation on the energy text by using a word segmentation toolkit, and then performing word deletion stopping operation on the text to reduce the influence of noise words on the classification effect;
(3) text vectorization, defining xiI-th word, x, representing a sequence of wordsi:jThe i to j words representing a sequence of words. Then the preprocessed Chinese word sequence x with the length of ni:nUsing Word2vec to directly train good Word vectors with the dimension of k, and splicing all the Word vectors to obtain a Word vector matrix with the dimension of n × k, wherein the Word vector matrix is used as the input of the convolutional layer;
(4) convolution operation, using a convolution kernel with a width equal to the dimension k of the word vector and a height h and a word window x containing h wordsi:i+h-1Convolution is carried out, and then a one-dimensional vector filter w with the length h × k acts on a word window to extract the characteristic ciThus, a n + h-1 dimensional characteristic diagram c ═ c is obtained1,c2,...,cn+h-1]；
(5) Performing pooling operation, performing maximal pooling operation on the feature map extracted by each filter, and obtaining a vector z with the length of m ═ z [ z ═ z ] through one-layer convolution and one-layer pooling for m filters1,z2,...,zm]；
(6) And (4) performing full-connection operation and classification, and feeding the vector z into a full-connection layer to obtain a final feature extraction vector. When classifying energy comment content, Softmax is used as an activation function, mainly classified into thermal power, hydroelectric power, wind power, tidal power, solar power, nuclear power, and hybrid energy. When energy comment emotions are classified, Sigmoid is used as an activation function and is mainly divided into negative, neutral and positive;
(7) l STM predictor model training, training network for energy trend prediction, training original time sequence T ═ T { [ T ]0,t1,...,tmDividing a training set and a test set to obtain Ttrain＝{t0,t1,...,tvAnd Ttest＝{tv,tv+1,...,tmWhen the time step is L, if the network input p is (p)1,p2,...,pL) Theoretically, the output q ═ q (q) is output1,q2,...,qL) And p passes through the hidden layer and outputs g ═ g (g)1,g2,...,gL) Then the loss function of the training process is:
(8) l STM predictor model prediction, when trend prediction is carried out, test set T is usedtest＝{tv,tv+1,...,tmInput p at the first time step inv＝(pv+1,pv+2,...,pv+L) Normalized and input to a trained L STM, the result is expressed as gv＝(gv+2,gv+3,...,gv+L+1) The predicted value obtained when m + L is gv+L+1. By analogy, the predicted time sequence is gtest＝(gv+1,gv+2,...,gm) To obtain gtestAnd then, carrying out reverse normalization on the time sequence to finally obtain a corresponding prediction time sequence.
By adopting the method for realizing the energy power public opinion analysis processing based on emotion mining, an analysis model aiming at the energy development trend of an energy report is realized, the model well evaluates the public opinion trend of a power source, has stronger topic analysis and sequence mining prediction capabilities, and can be expanded to other time sequence public opinion analysis and prediction fields. In addition, fine-grained optimization can be continuously carried out on the theme mode of the algorithm, multi-theme multi-sequence emotion comparison analysis is realized by analyzing the internal structure of the information news, more comprehensive public sentiment research on the power trend is developed, the research on the energy development trend can be mastered in time, the mutual evidence is obtained with the energy development quantitative analysis, the future energy development trend and law can be grasped more objectively and accurately, and theoretical guidance is provided for constructing an energy system.
In this specification, the invention has been described with reference to specific embodiments thereof. It will, however, be evident that various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention. The specification and drawings are, accordingly, to be regarded in an illustrative rather than a restrictive sense.
Claims (6)
1. A method for realizing energy power public opinion analysis processing based on emotion mining is characterized by comprising the following steps:
(1) collecting samples, and collecting information through a web crawler;
(2) performing text preprocessing, and segmenting words of the text through a word segmentation toolkit;
(3) performing text vectorization through word2vec technology;
(4) performing convolution operation;
(5) performing pooling operation;
(6) performing full-link operation, and performing energy theme classification and emotional tendency identification according to the convolutional neural network;
(7) and (4) carrying out energy public opinion tendency prediction through an L STM sequence mining network.
2. The method for realizing the energy power public opinion analysis processing based on emotion mining as claimed in claim 1, wherein the step (2) specifically comprises the following steps:
and performing word segmentation on the energy text through a word segmentation toolkit, and performing word deletion and stop operation on the text.
3. The method for realizing the energy power public opinion analysis processing based on emotion mining as claimed in claim 1, wherein the step (3) specifically comprises the following steps:
defining a word sequence, and splicing the trained word vectors to obtain a word vector matrix which is used as the input of the convolutional layer.
4. The method for realizing the energy power public opinion analysis based on emotion mining as claimed in claim 1, wherein the step (5) specifically comprises the following steps:
and performing maximum pooling operation on the feature map extracted by each filter, and obtaining a vector through one-layer convolution and one-layer pooling.
5. The method for realizing the energy power public opinion analysis based on emotion mining as claimed in claim 1, wherein the step (6) specifically comprises the following steps:
(6.1) feeding the vectors into a full connection layer to obtain final feature extraction vectors;
(6.2) judging whether the energy comment content is classified or the energy comment emotion is classified, and if the energy comment content is classified, taking a Softmax function as an activation function to classify the energy comment content into thermal power, hydroelectric power, wind power, tidal power, solar energy, nuclear energy and mixed energy; if the energy comment emotion is classified, the Sigmoid function is used as an activation function and is classified into negative, neutral and positive.
6. The method for realizing the energy power public opinion analysis based on emotion mining as claimed in claim 1, wherein the step (7) specifically comprises the following steps:
(7.1) training L STM predictor model;
(7.2) prediction was performed by L STM predictor model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010300741.7A CN111506793A (en) | 2020-04-16 | 2020-04-16 | Method for realizing energy power public opinion analysis processing based on emotion mining |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010300741.7A CN111506793A (en) | 2020-04-16 | 2020-04-16 | Method for realizing energy power public opinion analysis processing based on emotion mining |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111506793A true CN111506793A (en) | 2020-08-07 |
Family
ID=71877590
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010300741.7A Pending CN111506793A (en) | 2020-04-16 | 2020-04-16 | Method for realizing energy power public opinion analysis processing based on emotion mining |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111506793A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110083700A (en) * | 2019-03-19 | 2019-08-02 | 北京中兴通网络科技股份有限公司 | A kind of enterprise's public sentiment sensibility classification method and system based on convolutional neural networks |
| CN110162626A (en) * | 2019-04-26 | 2019-08-23 | 湘潭大学 | A kind of calculation method of the public sentiment emotion temperature entropy based on two-way LSTM |
| CN110188933A (en) * | 2019-05-21 | 2019-08-30 | 湖北经济学院 | A kind of School Network public sentiment monitoring and pre-warning method and system |
| CN110851594A (en) * | 2019-10-08 | 2020-02-28 | 浙江工业大学 | Text classification method and device based on multi-channel deep learning model |
-
2020
- 2020-04-16 CN CN202010300741.7A patent/CN111506793A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110083700A (en) * | 2019-03-19 | 2019-08-02 | 北京中兴通网络科技股份有限公司 | A kind of enterprise's public sentiment sensibility classification method and system based on convolutional neural networks |
| CN110162626A (en) * | 2019-04-26 | 2019-08-23 | 湘潭大学 | A kind of calculation method of the public sentiment emotion temperature entropy based on two-way LSTM |
| CN110188933A (en) * | 2019-05-21 | 2019-08-30 | 湖北经济学院 | A kind of School Network public sentiment monitoring and pre-warning method and system |
| CN110851594A (en) * | 2019-10-08 | 2020-02-28 | 浙江工业大学 | Text classification method and device based on multi-channel deep learning model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110598005B (en) | Public safety event-oriented multi-source heterogeneous data knowledge graph construction method | |
| CN107193801B (en) | Short text feature optimization and emotion analysis method based on deep belief network | |
| CN106886580B (en) | Image emotion polarity analysis method based on deep learning | |
| CN108090070B (en) | Chinese entity attribute extraction method | |
| CN112347268A (en) | Text-enhanced knowledge graph joint representation learning method and device | |
| CN110362819B (en) | Text emotion analysis method based on convolutional neural network | |
| CN111310476B (en) | Public opinion monitoring method and system using aspect-based emotion analysis method | |
| CN110750635B (en) | French recommendation method based on joint deep learning model | |
| CN107168956B (en) | Chinese chapter structure analysis method and system based on pipeline | |
| CN116304745B (en) | Text topic matching method and system based on deep semantic information | |
| CN112925904A (en) | Lightweight text classification method based on Tucker decomposition | |
| CN116842194A (en) | Electric power semantic knowledge graph system and method | |
| TW202111569A (en) | Text classification method with high scalability and multi-tag and apparatus thereof also providing a method and a device for constructing topic classification templates | |
| CN116451114A (en) | Internet of things enterprise classification system and method based on enterprise multisource entity characteristic information | |
| CN109299251A (en) | A kind of abnormal refuse messages recognition methods and system based on deep learning algorithm | |
| CN113609267B (en) | Speech relation recognition method and system based on GCNDT-MacBERT neural network framework | |
| CN106598952A (en) | System for detecting Chinese fuzzy constraint information scope based on convolutional neural network | |
| CN115995029A (en) | Image emotion analysis method based on bidirectional connection | |
| CN111506793A (en) | Method for realizing energy power public opinion analysis processing based on emotion mining | |
| CN113961708A (en) | Power equipment fault tracing method based on multilevel graph convolutional network | |
| CN114298041A (en) | Network security named entity identification method and identification device | |
| CN114064888A (en) | Financial text classification method and system based on BERT-CNN | |
| CN114491033A (en) | Method for building user interest model based on word vector and topic model | |
| Wang et al. | Exploring topic supervision with bert for text matching | |
| CN113076424A (en) | Data enhancement method and system for unbalanced text classified data |
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 |