CN119047486A - Financial expert language semantic emotion analysis system and method - Google Patents
Financial expert language semantic emotion analysis system and method Download PDFInfo
- Publication number
- CN119047486A CN119047486A CN202411523783.1A CN202411523783A CN119047486A CN 119047486 A CN119047486 A CN 119047486A CN 202411523783 A CN202411523783 A CN 202411523783A CN 119047486 A CN119047486 A CN 119047486A
- Authority
- CN
- China
- Prior art keywords
- financial
- market
- sentiment
- data
- speech data
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; 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
- G06F16/353—Clustering; Classification into predefined classes
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- G06N3/0455—Auto-encoder networks; Encoder-decoder networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/06—Asset management; Financial planning or analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Business, Economics & Management (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Development Economics (AREA)
- Databases & Information Systems (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Strategic Management (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Human Resources & Organizations (AREA)
- Technology Law (AREA)
- Machine Translation (AREA)
Abstract
本发明涉及金融数据处理技术领域，具体涉及一种金融专家语言语义感情分析系统及方法；本发明通过对金融专家言论数据进行分词、词性标注、实体识别和专业术语提取，形成词嵌入矩阵，对词嵌入矩阵中每个文本向量进行情感极性分析，并针对金融专家言论数据进行类别分类，结合情感极性和类别分类得到词嵌入矩阵的情感强度，基于历史金融专家言论数据和其对应的历史市场反馈数据构建市场情绪评估模型，采用市场情绪评估模型评估词嵌入矩阵的市场情绪，并将市场情绪与市场关键指标进行相关性分析，得到金融专家言论数据对市场关键指标的影响；通过本发明实现对金融专家言论数据的情感分析，实现金融专家言论数据对市场关键指标的影响分析。
The present invention relates to the technical field of financial data processing, and in particular to a financial expert language semantic sentiment analysis system and method; the present invention forms a word embedding matrix by performing word segmentation, part-of-speech tagging, entity recognition and professional terminology extraction on financial expert speech data, performs sentiment polarity analysis on each text vector in the word embedding matrix, and performs category classification on the financial expert speech data, obtains the sentiment intensity of the word embedding matrix by combining the sentiment polarity and category classification, constructs a market sentiment evaluation model based on historical financial expert speech data and its corresponding historical market feedback data, uses the market sentiment evaluation model to evaluate the market sentiment of the word embedding matrix, and performs correlation analysis between the market sentiment and market key indicators to obtain the influence of the financial expert speech data on the market key indicators; the present invention realizes the sentiment analysis of the financial expert speech data and the influence analysis of the financial expert speech data on the market key indicators.
Description
Technical Field
The invention relates to the technical field of financial data processing, in particular to a financial expert language semantic emotion analysis system and a financial expert language semantic emotion analysis method.
Background
In the current financial field, analysis of financial expert language mainly depends on manual interpretation, and manual interpretation also depends on financial professionals, so that interpretation efficiency is low and interpretation conclusion is easily influenced by subjective factors. Existing automated analysis tools often fail to accurately understand financial terms and complex contexts, resulting in inaccurate emotion analysis. In addition, the lack of predictive and risk pre-warning functions for market emotional trends makes it difficult for investors and decision makers to grasp market dynamics and potential risks.
Disclosure of Invention
In view of the above-mentioned shortcomings of the prior art, the present invention aims to provide a system and a method for analyzing the semantic emotion of a financial expert language, so as to solve the problem that it is difficult to obtain effective and reliable financial information from the financial expert language.
In order to solve the problems, the invention adopts the following technical scheme:
In one aspect, the invention provides a financial expert language semantic emotion analysis system, which comprises a natural language processing module, an emotion analysis module and a market emotion assessment module;
The natural language processing module is used for carrying out word segmentation, part-of-speech tagging, entity identification and professional term extraction on financial expert language data to form a word embedding matrix;
The emotion analysis module is used for carrying out emotion polarity analysis on each text vector in the word embedding matrix, wherein the emotion polarities comprise positive, negative and neutral, and carrying out optimistic, cautious or pessimistic category classification on financial expert speaking data, and combining the emotion polarities and the category classification to obtain the emotion intensity of the word embedding matrix;
The market emotion assessment module is used for constructing a market emotion assessment model based on historical financial expert speaking data and corresponding historical market feedback data, assessing the market emotion of the word embedded matrix by adopting the market emotion assessment model, and carrying out correlation analysis on the market emotion and the market key index to obtain the influence of the financial expert speaking data on the market key index.
As an implementation manner, the word segmentation, part-of-speech tagging, entity identification and term extraction are performed on financial expert speech data to form a word embedding matrix, which includes:
word segmentation is carried out on financial expert speaking data by adopting a word segmentation model trained by a corpus in the financial field;
Labeling the parts of speech of the segmented financial expert speaking data;
Identifying financial entities of the financial expert speaking data after part of speech tagging by adopting a long-short-term memory network or a transducer model;
extracting the technical terms by combining a deep learning model of a dictionary in the financial field to form a financial technical term library, and extracting the technical terms of the financial entity by calculating cosine similarity of the identified financial entity and the technical terms in the financial technical term library;
A word embedding matrix of financial expert speech data is constructed based on the extracted technical terms of the financial entity.
As an implementation manner, the emotion polarity analysis on each text vector in the word embedding matrix includes:
analyzing each text vector in the word embedding matrix by using the trained deep neural network model, and calculating the emotion polarity score of each word;
The optimistic, cautious or pessimistic class classification of financial expert speech data includes:
Converting the context information of the financial expert language data into a feature vector by using TF-IDF or word2vec, and integrating the feature vector with the score of the emotion polarity to obtain a comprehensive feature vector;
classifying the financial expert speaking data by a trained softmax regression model;
The emotion intensity of the word embedding matrix obtained by combining emotion polarity and category classification comprises the following steps:
and calculating the emotion intensity of the word embedding matrix according to the weight and polarity scores of different emotion categories.
As an implementation manner, the building the market emotion assessment model based on the historical financial expert language data and the corresponding historical market feedback data includes:
collecting historical language data and related market feedback data of financial experts, and preprocessing the data;
Extracting key features from the preprocessed data to construct key feature vectors, wherein the key features comprise scores of emotion polarities of historical language data of financial experts, category classification and market key indexes of related market feedback data;
And constructing realization sequence data based on the constructed key feature vector, and training by adopting a long-short-term memory network to obtain a market emotion assessment model.
As an implementation manner, the estimating the market emotion of the word embedding matrix by using the market emotion estimation model, and performing correlation analysis on the market emotion and the market key index to obtain the influence of the financial expert speaking data on the market key index includes:
And obtaining the influence of financial expert speaking data on the market key indexes according to the obtained market emotion of the word embedding matrix, the score of emotion polarity and the change curve of the market key indexes of the market feedback data related to category classification matching.
In another aspect, the present invention provides a financial expert language semantic emotion analysis method, including:
word segmentation, part-of-speech tagging, entity identification and professional term extraction are carried out on financial expert speaking data to form a word embedding matrix;
carrying out emotion polarity analysis on each text vector in the word embedding matrix, wherein the emotion polarities comprise positive, negative and neutral, and carrying out optimistic, cautious or pessimistic category classification on financial expert speaking data, and combining the emotion polarities and the category classification to obtain the emotion intensity of the word embedding matrix;
And constructing a market emotion assessment model based on the historical financial expert speaking data and the corresponding historical market feedback data, assessing the market emotion of the word embedded matrix by adopting the market emotion assessment model, and carrying out correlation analysis on the market emotion and the market key index to obtain the influence of the financial expert speaking data on the market key index.
As an implementation manner, the word segmentation, part-of-speech tagging, entity identification and term extraction are performed on financial expert speech data to form a word embedding matrix, which includes:
word segmentation is carried out on financial expert speaking data by adopting a word segmentation model trained by a corpus in the financial field;
Labeling the parts of speech of the segmented financial expert speaking data;
Identifying financial entities of the financial expert speaking data after part of speech tagging by adopting a long-short-term memory network or a transducer model;
extracting the technical terms by combining a deep learning model of a dictionary in the financial field to form a financial technical term library, and extracting the technical terms of the financial entity by calculating cosine similarity of the identified financial entity and the technical terms in the financial technical term library;
A word embedding matrix of financial expert speech data is constructed based on the extracted technical terms of the financial entity.
As an implementation manner, the emotion polarity analysis on each text vector in the word embedding matrix includes:
analyzing each text vector in the word embedding matrix by using the trained deep neural network model, and calculating the emotion polarity score of each word;
The optimistic, cautious or pessimistic class classification of financial expert speech data includes:
Converting the context information of the financial expert language data into a feature vector by using TF-IDF or word2vec, and integrating the feature vector with the score of the emotion polarity to obtain a comprehensive feature vector;
classifying the financial expert speaking data by a trained softmax regression model;
The emotion intensity of the word embedding matrix obtained by combining emotion polarity and category classification comprises the following steps:
and calculating the emotion intensity of the word embedding matrix according to the weight and polarity scores of different emotion categories.
As an implementation manner, the building the market emotion assessment model based on the historical financial expert language data and the corresponding historical market feedback data includes:
collecting historical language data and related market feedback data of financial experts, and preprocessing the data;
Extracting key features from the preprocessed data to construct key feature vectors, wherein the key features comprise scores of emotion polarities of historical language data of financial experts, category classification and market key indexes of related market feedback data;
And constructing realization sequence data based on the constructed key feature vector, and training by adopting a long-short-term memory network to obtain a market emotion assessment model.
As an implementation manner, the estimating the market emotion of the word embedding matrix by using the market emotion estimation model, and performing correlation analysis on the market emotion and the market key index to obtain the influence of the financial expert speaking data on the market key index includes:
And obtaining the influence of financial expert speaking data on the market key indexes according to the obtained market emotion of the word embedding matrix, the score of emotion polarity and the change curve of the market key indexes of the market feedback data related to category classification matching.
The financial expert language semantic emotion analysis system and the financial expert language semantic emotion analysis method have the beneficial effects that emotion polarity analysis and category classification are carried out on financial expert language data, emotion analysis on the financial expert language data is achieved, correlation is carried out on the basis of market feedback data, the relation between the financial expert language data and market key indexes is obtained, and therefore influence analysis of the financial expert language data on the market key indexes is achieved.
Drawings
FIG. 1 is a schematic diagram of a financial expert language semantic emotion analysis system according to the present invention.
FIG. 2 is a flow chart of a method for analyzing the semantic emotion of a financial expert language.
Detailed Description
The present invention will be described in further detail with reference to specific examples.
It should be noted that these examples are only for illustrating the present invention, and not for limiting the present invention, and simple modifications of the method under the premise of the inventive concept are all within the scope of the claimed invention.
Referring to fig. 1, a financial expert language semantic emotion analysis system includes a natural language processing module 100, an emotion analysis module 200, and a market emotion assessment module 300.
The natural language processing module 100 is used for word segmentation, part-of-speech tagging, entity identification and term extraction on financial expert speech data to form a word embedding matrix.
The word segmentation of the financial expert speaking data comprises word segmentation of the financial expert speaking data by adopting a word segmentation model trained by a corpus in the financial field. The word segmentation model can be combined with a rule method and a statistical method, so that common words can be recognized, and financial professional terms can be accurately segmented.
The part-of-speech tagging comprises the steps of tagging part-of-speech of financial expert after word segmentation, and assigning part-of-speech tags such as nouns, verbs, adjectives and the like to each vocabulary so as to provide more accurate semantic information for subsequent emotion analysis.
The entity identification comprises the step of identifying financial entities of the word-part tagged financial expert speaking data by adopting a long-short-term memory network or a transducer model. Financial entities in the text, such as person names, institution names, stock codes, etc., are identified to facilitate extraction of key information.
The term of art extraction includes that a deep learning model combined with a dictionary in the financial field is adopted to extract terms of art to form a financial term library, and the cosine similarity of the identified financial entity and the terms in the financial term library is calculated to extract the terms of the financial entity.
Forming the word embedding matrix includes constructing the word embedding matrix of financial expert speech data based on the extracted terminology of the financial entity.
Wherein, The word is represented as being embedded in a matrix,The sequence of words to be input is represented,Is a word embedding function that maps each word into a dense vector of fixed dimensions that captures the semantic information of the word.
For example, consider a section of financial expert comments that "recent market fluctuations are large, but the investment value of high-quality stocks is still significant". In the implementation process, sentences are firstly divided into word units such as 'recent', 'market', 'fluctuation', 'larger' and the like through word segmentation processing. Subsequently, part-of-speech tagging is performed, such as "recent" being noted as a temporal noun and "larger" being noted as an adjective. The entity identification and term extraction step will then identify financial entities and terms such as "market" and "premium stocks". Finally, through word embedding technology, each vocabulary is converted into a vector capable of expressing the deep semantic meaning of the vocabulary, and the vectors are then input into a deep learning model for emotion analysis to judge the emotion tendency expressed by the comment.
The emotion analysis module 200 is configured to analyze emotion polarity of each text vector in the word embedding matrix, where the emotion polarity includes positive, negative and neutral, and performs optimistic, cautious or pessimistic category classification on financial expert speech data, and combines the emotion polarity and the category classification to obtain emotion intensity of the word embedding matrix.
The process specifically comprises the following steps:
And analyzing each text vector in the word embedding matrix by using the trained deep neural network model, and calculating the emotion polarity score of each word. The deep neural network may be a Convolutional Neural Network (CNN) or a Recurrent Neural Network (RNN).
Wherein, A score representing the polarity of the emotion,For the word to be embedded in the matrix,Is an emotion polarity analysis function.
The optimistic, cautious or pessimistic classification of category for financial expert speech data includes:
And converting the context information of the financial expert language data into a feature vector by using TF-IDF or word2vec, and integrating the feature vector with the score of the emotion polarity to obtain a comprehensive feature vector. The context information includes:
The mood level, the context information is not divided into several levels, but is obtained and expressed in various ways, and the following are several common obtaining ways:
lexical characteristics, such as part of speech, word frequency, word embedding (word embeddings), etc.
Sentence characteristics, such as sentence length, sentence structure, grammar analysis, etc.
The chapter features are chapter structures, paragraph relations, topic consistency and the like.
The financial expert speech data was category-classified by a trained softmax regression model.
The softmax regression model is trained using comprehensive feature vectors constructed from emotional and contextual features extracted from financial expert speech data, and outputs classification probabilities that the financial expert speech data belongs to optimistic, cautious or pessimistic.
Wherein, A probability distribution representing the classification of the text,The context information is represented by a representation of the context information,A score representing the polarity of the emotion,Is a text classification function.
The method for obtaining the emotion strength of the word embedding matrix by combining emotion polarity and category classification comprises the following steps:
and calculating the emotion intensity of the word embedding matrix according to the weight and polarity scores of different emotion categories.
Wherein I represents an emotion intensity value,Is the weight of the i-th emotion class,Is the score of emotion polarity, N is the total number of words.
For example, the word "wave" may be determined as neutral, while "premium stock" may be determined as positive emotion. In the text classification phase, the entire comment may be divided into "optimistic" categories. In the emotional intensity assessment, the comment may obtain a higher emotional intensity value according to the model calculation, indicating that the comment has a larger positive influence on the market emotion. Through this series of analyses, the present invention is able to provide investors with quantitative assessments concerning market emotion, thereby assisting investment decisions.
Through the steps, the emotion tendencies of financial language can be accurately judged, emotion intensity evaluation can be carried out by combining with context factors, and an efficient and reliable solution is provided for market emotion analysis.
The market emotion assessment module 300 is configured to construct a market emotion assessment model based on historical financial expert speech data and corresponding historical market feedback data, assess the market emotion of the word embedded matrix by adopting the market emotion assessment model, and perform correlation analysis on the market emotion and the market key index to obtain the influence of the financial expert speech data on the market key index.
Wherein, the constructing of the market emotion estimation model includes:
historical language data and related market feedback data of financial experts are collected and data preprocessing is performed.
The pretreatment comprises the following steps:
Historical speech data and market feedback data for financial professionals are collected and consolidated, including but not limited to stock price fluctuations, trading volume changes, social media moods, and the like. And cleaning and normalizing the data to eliminate noise and abnormal values in the data.
Wherein, Is the data after the normalization and is used for the data,Is the original data of the data set,AndRespectively minimum and maximum values in the dataset.
And extracting key features from the preprocessed data to construct key feature vectors, wherein the key features comprise scores of emotion polarities of historical language data of financial experts, category classification and market key indexes of related market feedback data.
Wherein, Is a feature vector of the object set,Is the i-th feature.
And constructing realization sequence data based on the constructed key feature vector, and training by adopting a long-short-term memory network to obtain a market emotion assessment model.
Wherein, Is the result of the assessment of the emotion of the market,Is a model parameter.
Evaluating the market emotion of the word embedding matrix by adopting a market emotion evaluation model, and performing correlation analysis on the market emotion and the market key index, wherein the obtaining of the influence of financial expert speaking data on the market key index comprises the following steps:
And obtaining the influence of financial expert speaking data on the market key indexes according to the obtained market emotion of the word embedding matrix and the change curve of the market key indexes (such as index rise and fall and transaction amount change) of the market feedback data related to the classification matching of the emotion polarity and the category.
For example, historical speech data and related market feedback data of the expert are collected first through a data preprocessing step. In the feature extraction stage, the emotion polarity score of the comment, the text classification result and the like are taken as features. And then, processing the feature vector by using the constructed LSTM emotion estimation model to obtain a market emotion estimation result. Through emotion quantitative analysis, a remarkable positive correlation exists between the speech of the expert and the market transaction amount, so that the speech of the expert has a pushing effect on market transaction liveness and/or the speech of the expert has market correctness. This analysis can provide investors with deep insight into market dynamics, assisting them in making more reasonable investment decisions.
Referring to fig. 2, a method for analyzing semantic emotion of financial expert language includes:
S100, word segmentation, part-of-speech tagging, entity identification and professional term extraction are carried out on financial expert speaking data to form a word embedding matrix;
S200, carrying out emotion polarity analysis on each text vector in the word embedding matrix, wherein the emotion polarities comprise positive, negative and neutral, and carrying out optimistic, cautious or pessimistic category classification on financial expert speaking data, and obtaining the emotion strength of the word embedding matrix by combining the emotion polarities and the category classification;
S300, a market emotion assessment model is built based on historical financial expert speaking data and corresponding historical market feedback data, the market emotion of the word embedding matrix is assessed by adopting the market emotion assessment model, and correlation analysis is carried out on the market emotion and the market key index, so that influence of the financial expert speaking data on the market key index is obtained.
The word embedding matrix is formed by word segmentation, part-of-speech tagging, entity identification and professional term extraction of financial expert speaking data, and comprises the following steps:
word segmentation is carried out on financial expert speaking data by adopting a word segmentation model trained by a corpus in the financial field;
Labeling the parts of speech of the segmented financial expert speaking data;
Identifying financial entities of the financial expert speaking data after part of speech tagging by adopting a long-short-term memory network or a transducer model;
extracting the technical terms by combining a deep learning model of a dictionary in the financial field to form a financial technical term library, and extracting the technical terms of the financial entity by calculating cosine similarity of the identified financial entity and the technical terms in the financial technical term library;
A word embedding matrix of financial expert speech data is constructed based on the extracted technical terms of the financial entity.
The emotion polarity analysis for each text vector in the word embedding matrix comprises the following steps:
analyzing each text vector in the word embedding matrix by using the trained deep neural network model, and calculating the emotion polarity score of each word;
The optimistic, cautious or pessimistic class classification of financial expert speech data includes:
Converting the context information of the financial expert language data into a feature vector by using TF-IDF or word2vec, and integrating the feature vector with the score of the emotion polarity to obtain a comprehensive feature vector;
classifying the financial expert speaking data by a trained softmax regression model;
The emotion intensity of the word embedding matrix obtained by combining emotion polarity and category classification comprises the following steps:
and calculating the emotion intensity of the word embedding matrix according to the weight and polarity scores of different emotion categories.
Wherein the constructing the market emotion assessment model based on the historical financial expert speech data and the corresponding historical market feedback data comprises:
collecting historical language data and related market feedback data of financial experts, and preprocessing the data;
Extracting key features from the preprocessed data to construct key feature vectors, wherein the key features comprise scores of emotion polarities of historical language data of financial experts, category classification and market key indexes of related market feedback data;
And constructing realization sequence data based on the constructed key feature vector, and training by adopting a long-short-term memory network to obtain a market emotion assessment model.
The method for estimating the market emotion of the word embedding matrix by adopting the market emotion estimation model, and carrying out correlation analysis on the market emotion and the market key index, wherein the step of obtaining the influence of financial expert speaking data on the market key index comprises the following steps:
And obtaining the influence of financial expert speaking data on the market key indexes according to the obtained market emotion of the word embedding matrix, the score of emotion polarity and the change curve of the market key indexes of the market feedback data related to category classification matching.
Finally, it is noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be understood that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411523783.1A CN119047486B (en) | 2024-10-30 | 2024-10-30 | A financial expert language semantic sentiment analysis system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411523783.1A CN119047486B (en) | 2024-10-30 | 2024-10-30 | A financial expert language semantic sentiment analysis system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119047486A true CN119047486A (en) | 2024-11-29 |
| CN119047486B CN119047486B (en) | 2025-02-11 |
Family
ID=93569113
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411523783.1A Active CN119047486B (en) | 2024-10-30 | 2024-10-30 | A financial expert language semantic sentiment analysis system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119047486B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119293251A (en) * | 2024-12-09 | 2025-01-10 | 江苏苏商银行股份有限公司 | A method for analyzing financial user sentiment and psychological portraits based on a large language model |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150286710A1 (en) * | 2014-04-03 | 2015-10-08 | Adobe Systems Incorporated | Contextualized sentiment text analysis vocabulary generation |
| CN106919673A (en) * | 2017-02-21 | 2017-07-04 | 浙江工商大学 | Text mood analysis system based on deep learning |
| CN109933664A (en) * | 2019-03-12 | 2019-06-25 | 中南大学 | An Improved Method for Fine-Grained Sentiment Analysis Based on Sentiment Word Embedding |
| CN109934503A (en) * | 2019-03-19 | 2019-06-25 | 合肥工业大学 | An early warning method of financial market risk in the Internet environment |
| WO2024021354A1 (en) * | 2022-07-28 | 2024-02-01 | 中国科学院深圳先进技术研究院 | Model training method, price prediction method, terminal device and storage medium |
| CN118628113A (en) * | 2024-06-24 | 2024-09-10 | 天元大数据信用管理有限公司 | A financial transaction risk assessment method, device and medium based on machine learning |
-
2024
- 2024-10-30 CN CN202411523783.1A patent/CN119047486B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150286710A1 (en) * | 2014-04-03 | 2015-10-08 | Adobe Systems Incorporated | Contextualized sentiment text analysis vocabulary generation |
| CN106919673A (en) * | 2017-02-21 | 2017-07-04 | 浙江工商大学 | Text mood analysis system based on deep learning |
| CN109933664A (en) * | 2019-03-12 | 2019-06-25 | 中南大学 | An Improved Method for Fine-Grained Sentiment Analysis Based on Sentiment Word Embedding |
| CN109934503A (en) * | 2019-03-19 | 2019-06-25 | 合肥工业大学 | An early warning method of financial market risk in the Internet environment |
| WO2024021354A1 (en) * | 2022-07-28 | 2024-02-01 | 中国科学院深圳先进技术研究院 | Model training method, price prediction method, terminal device and storage medium |
| CN118628113A (en) * | 2024-06-24 | 2024-09-10 | 天元大数据信用管理有限公司 | A financial transaction risk assessment method, device and medium based on machine learning |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119293251A (en) * | 2024-12-09 | 2025-01-10 | 江苏苏商银行股份有限公司 | A method for analyzing financial user sentiment and psychological portraits based on a large language model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119047486B (en) | 2025-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | A deep learning-based approach to constructing a domain sentiment lexicon: a case study in financial distress prediction | |
| US11886814B2 (en) | Systems and methods for deviation detection, information extraction and obligation deviation detection | |
| CN108563638B (en) | Microblog emotion analysis method based on topic identification and integrated learning | |
| Jiang et al. | Mining semantic features in current reports for financial distress prediction: Empirical evidence from unlisted public firms in China | |
| CN110347787B (en) | Interview method and device based on AI auxiliary interview scene and terminal equipment | |
| CN118821045B (en) | Knowledge-enhanced product question-answering community user conversation emotion recognition method and system | |
| CN113111152A (en) | Depression detection method based on knowledge distillation and emotion integration model | |
| CN116610592B (en) | Customizable software testing and evaluation method and system based on natural language processing technology | |
| CN109726745A (en) | A goal-based sentiment classification method incorporating descriptive knowledge | |
| CN119047486B (en) | A financial expert language semantic sentiment analysis system and method | |
| CN119577522B (en) | Method for correcting old man dialogue intention information based on deep neural network | |
| CN118171645B (en) | Business information analysis method and system based on text classification | |
| CN111651606A (en) | Text processing method and device and electronic equipment | |
| CN111159405B (en) | Irony detection method based on background knowledge | |
| CN111145903A (en) | Method and device for acquiring vertigo inquiry text, electronic equipment and inquiry system | |
| CN118193731A (en) | Method and system for topic identification and clustering screening of scientific and technological texts based on SAO structure | |
| CN118013975A (en) | Natural language processing method and system based on artificial intelligence | |
| CN119441389A (en) | A financial sentiment analysis system | |
| Jui et al. | A machine learning-based segmentation approach for measuring similarity between sign languages | |
| CN119538919B (en) | Text feature data processing method and system based on BERT model and TF-IDF weighting | |
| CN106227802A (en) | A kind of based on Chinese natural language process and the multiple source Forecasting of Stock Prices method of multi-core classifier | |
| Korade et al. | Elevating intelligent voice assistant chatbots with natural language processing, and OpenAI technologies | |
| CN120015307A (en) | An intelligent method and system for rapid screening and triage of aortic dissection | |
| CN117788137A (en) | Prediction model and construction method of credit default of small and medium-sized enterprises based on unstructured data | |
| CN116340521A (en) | Document classification method and device, storage medium and electronic equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |