A feedback information hazard degree analysis system and method based on artificial intelligence

Abstract

The present invention relates to the field of network information processing technology, and discloses a feedback information hazard degree analysis system and method based on artificial intelligence. The feedback information hazard degree analysis system based on artificial intelligence includes a data acquisition module, a text data analysis module, a hazard degree assessment module, an early warning response module, and a visual display module; the hazard degree assessment module is used to customize the feedback information hazard degree assessment for different users by combining the emotional tendency of financial feedback information with user characteristics, and predict the impact of feedback information on financial market behavior in combination with historical data. The present invention can accurately analyze the emotional tendency and user characteristics of feedback information, and provide customized hazard degree assessment based on these analyses, thereby providing a powerful data-driven decision support tool for financial market participants, helping them to better understand market dynamics and make more informed decisions.

Description

Technical Field

[0001] The present invention relates to the technical field of network information processing, and in particular to an artificial intelligence-based feedback information hazard degree analysis system and method. Background Art

[0002] Feedback information system is a technical system that uses artificial intelligence and big data technology to analyze and monitor public opinion. It collects, organizes and analyzes data from various public information sources such as social media, news, forums, Weibo, etc. to understand and evaluate the public's attitude and emotions towards specific topics, events or brands. With the popularity of the Internet and social media, public opinion has become increasingly important to governments, businesses and individuals. The development of feedback information systems aims to help people better understand and respond to public opinion, as well as to promptly detect and resolve potential crises and problems.

[0003] The development of feedback information systems has gone through the following stages: 1) Data collection and organization: Early feedback information systems focused on data collection and organization, by grabbing and crawling information on the Internet and organizing it into structured data for subsequent analysis. 2) Sentiment analysis and emotion recognition: With the development of natural language processing and machine learning technology, feedback information systems began to focus on sentiment analysis and emotion recognition. It can automatically identify and classify sentiment tendencies in texts, such as positive, negative or neutral, to help people understand the public's emotional attitudes towards specific topics. 3) Feedback information monitoring and early warning: Modern feedback information systems can not only analyze public feedback information, but also monitor and warn of feedback information changes in real time. It can provide timely feedback information dynamics through real-time data updates and automatic alarm mechanisms, enabling people to quickly respond to and handle feedback information events. 4) Feedback information analysis and decision support: The latest generation of feedback information systems can not only provide data and alarms, but also conduct in-depth feedback information analysis and provide decision support. It can reveal the laws and trends behind feedback information through data mining and machine learning algorithms, and provide more scientific decision-making basis for governments, enterprises and individuals.

[0004] Natural Language Processing (NLP) system is a technical system that uses artificial intelligence and computer science technology to process and analyze human language. NLP system aims to enable computers to understand, interpret and generate natural language so that they can better interact and communicate with humans. The development of NLP system has gone through multiple stages and technological breakthroughs. The earliest NLP systems were mainly based on rules and grammar, parsing and understanding text by writing complex rules. However, this method often has difficulty in handling complex language structures and semantics. With the development of machine learning and deep learning technologies, NLP systems have gradually turned to data-driven methods. These systems are trained using large amounts of text data to learn the patterns and regularities of language. Among them, technologies such as word embedding technology (Word Embedding) and recurrent neural networks (RNN) are widely used in tasks such as semantic understanding, sentiment analysis, and machine translation.

[0005] In recent years, the emergence of pre-trained models such as BERT and GPT has further promoted the development of NLP systems. These models can provide better semantic understanding and generation capabilities by pre-training on large-scale text data, thereby achieving better performance in various text processing tasks. However, although the existing NLP models can meet specific needs, such as sentiment analysis and correlation analysis, the false positive rate and missed positive rate are too high, and they cannot meet the needs of users in the financial market and investor relations management industry. Therefore, the present invention proposes a feedback information hazard analysis system and method based on artificial intelligence. Summary of the invention

[0006] In view of the problems in the related technology, the present invention proposes a feedback information hazard level analysis system and method based on artificial intelligence to overcome the above-mentioned technical problems existing in the existing related technology.

[0007] To this end, the specific technical solution adopted by the present invention is as follows:

[0008] According to one aspect of the present invention, there is provided a feedback information hazard degree analysis system based on artificial intelligence, the feedback information hazard degree analysis system based on artificial intelligence is used for financial feedback information, and the feedback information hazard degree analysis system based on artificial intelligence includes a data acquisition module, a text data analysis module, a hazard degree assessment module, an early warning response module and a visual display module;

[0009] The data collection module is used to obtain real-time online financial feedback information data and user behavior data including websites, forums, and social platforms;

[0010] The text data analysis module is used to analyze the acquired real-time network financial feedback information data and user behavior data to obtain the emotional tendency and user characteristics of the financial feedback information;

[0011] The harm level assessment module is used to conduct customized feedback information harm level assessment for different users based on the emotional tendency of financial feedback information and user characteristics, and to predict the impact of feedback information on financial market behavior based on historical data;

[0012] The warning response module is used to monitor the financial market and feedback information dynamics in real time, trigger a warning when the degree of harm reaches a preset threshold, and provide corresponding response suggestions;

[0013] The visualization display module is used to display financial market trends, feedback information fluctuations, hazard level assessment results and corresponding response suggestions, and generate a comprehensive report.

[0014] Preferably, the data acquisition module includes a financial data acquisition module and a user behavior data acquisition module;

[0015] The financial data acquisition module is used to use crawler technology to acquire real-time online financial feedback information data including websites, forums, and social platforms, and to perform correlation analysis on the real-time online financial feedback information data, and to eliminate data with a correlation lower than a preset correlation threshold according to the analysis results;

[0016] The user behavior data acquisition module is used to use crawler technology to acquire the user's characteristic data and historical investment data, the characteristic data includes the user's age information, gender information and occupation information, and the historical investment data includes the user's historical transaction data and preference settings.

[0017] Preferably, the text data analysis module includes a sentiment analysis module and a user feature analysis module;

[0018] The sentiment analysis module is used to identify the sentiment in the financial feedback information data through sentiment analysis technology, and the sentiment includes positive, neutral and negative;

[0019] The user characteristic analysis module is used to analyze the user's characteristic data and historical investment data to obtain the user's risk preference and investment strategy.

[0020] Preferably, the hazard level assessment module includes a customized feedback information hazard level analysis module and a financial market impact prediction module;

[0021] The customized feedback information hazard level analysis module is used to conduct customized feedback information hazard level assessment for different users based on the emotional tendency of financial feedback information combined with user characteristics, trust, interest similarity, risk preference and investment strategy;

[0022] The financial market impact prediction module is used to predict the impact of feedback information on financial market behavior through time series analysis and prediction models combined with historical data.

[0023] Preferably, the customized feedback information hazard level analysis module includes a user similarity calculation module, a similar user determination module, a feedback information similarity calculation module, a similar feedback information determination module, and a hazard level determination module;

[0024] The user similarity calculation module is used to calculate the comprehensive user similarity between the target user and other users in the historical feedback information database, and to sort the comprehensive user similarities;

[0025] The similar user determination module is used to obtain the user most similar to the target user according to the ranking result of the user comprehensive similarity, and analyze and obtain the harm degree of the first initial feedback information corresponding to the most similar user;

[0026] The feedback information similarity calculation module is used to calculate the comprehensive similarity of feedback information between the target feedback information and other feedback information in the historical feedback information database, and to sort the comprehensive similarity of the feedback information;

[0027] The similar feedback information determination module is used to obtain the feedback information most similar to the target feedback information according to the sorting result of the comprehensive similarity of the feedback information, and analyze and obtain the harm degree of the second initial feedback information corresponding to the most similar feedback information;

[0028] The harm level determination module is used to calculate the final feedback information harm level based on the first initial feedback information harm level and the second initial feedback information harm level in combination with a preset weight;

[0029] The calculation formula of the final feedback information hazard level is: final feedback information hazard level = U*first initial feedback information hazard level + (1-U)*second initial feedback information hazard level, where U represents the weight of the first initial feedback information hazard level, and the value range of U is [0.1,0.4].

[0030] Preferably, the user similarity calculation module calculates the user comprehensive similarities between the target user and other users in the historical feedback information database and sorts the user comprehensive similarities, including:

[0031] Calculate the age similarity, gender similarity, occupation similarity, trust and interest similarity between the target user and other users in the historical feedback information database respectively, and calculate the user similarity based on the age similarity, gender similarity, occupation similarity, trust and interest similarity;

[0032] The risk preference similarity and investment strategy similarity between the target user and other users in the historical feedback information database are calculated separately, and the user comprehensive similarity is obtained by combining the user similarity calculation. The user comprehensive similarity is sorted in descending order according to the similarity value.

[0033] The calculation formula of the user comprehensive similarity is:

[0034] S 综合 =ω1·S 用户 +ω2·S 偏好 +ω3·S 策略

[0035]

[0036]

[0037] In the formula, S 综合 represents the comprehensive similarity of users, S 用户 represents user similarity, ω1 represents the weight of user similarity, S 偏好 represents the risk preference similarity, ω2 represents the weight of risk preference, S 策略 represents the similarity of investment strategies, ω3 represents the weight of investment strategies; X (m,n) Indicates the similarity of user interests, D (m,n) represents the user feature similarity, T (m,n) Indicates user trust; R m represents the risk preference of user m, R n represents the risk preference of user n, R represents the maximum possible risk preference difference, I represents the number of investment strategy parameters, O mi represents the i-th investment strategy parameter of user m, O ni represents the i-th investment strategy parameter of user n, and Z represents the maximum possible strategy difference.

[0038] Preferably, the feedback information similarity calculation module calculates the comprehensive similarity of feedback information between the target feedback information and other feedback information in the historical feedback information database, and sorts the comprehensive similarities of the feedback information, including:

[0039] The sentiment tendency similarity, influence range similarity and propagation speed similarity between the target feedback information and other feedback information in the historical feedback information database are calculated respectively, and the comprehensive similarity of the feedback information is obtained based on the sentiment tendency similarity, influence range similarity and propagation speed similarity, and the comprehensive similarity of the feedback information is sorted in descending order according to the similarity value.

[0040] The calculation formula of the comprehensive similarity of the feedback information is:

[0041] Y 综合 =ε1·Y 情感 +ε2·Y 影响 +ε3·Y 传播

[0042]

[0043] Where Y 综合 represents the comprehensive similarity of feedback information, Y 情感 represents the similarity of sentiment tendency, ε1 represents the weight of sentiment tendency, Y 影响 represents the similarity of the influence range, ε2 represents the weight of the influence range, and Y 传播 represents the similarity of propagation speed, ε3 represents the weight of propagation speed; E A Indicates the emotional tendency of feedback information A, E B represents the emotional tendency of the feedback information B, E represents the maximum emotional tendency difference, which is determined according to the quantitative range of the emotional tendency, and W A represents the impact range of feedback information A, W B represents the influence range of feedback information B, W represents the maximum influence range difference, which depends on the maximum and minimum influence ranges, V A represents the propagation speed of feedback information A, V B represents the propagation speed of the feedback information B, and V represents the maximum propagation speed difference, which depends on the fastest and slowest propagation speeds.

[0044] According to another aspect of the present invention, a method for analyzing the degree of harm of feedback information based on artificial intelligence is provided, and the method for analyzing the degree of harm of feedback information based on artificial intelligence comprises the following steps:

[0045] S1. Obtain real-time online financial feedback information data and user behavior data including websites, forums, and social platforms;

[0046] S2. Analyze the acquired real-time online financial feedback information data and user behavior data to obtain the emotional tendency and user characteristics of the financial feedback information;

[0047] S3. Use the sentiment tendency of financial feedback information combined with user characteristics to conduct customized feedback information harm assessment for different users, and combine historical data to predict the impact of feedback information on financial market behavior;

[0048] S4. Real-time monitoring of financial markets and feedback information dynamics, triggering warnings when the degree of harm reaches a preset threshold, and providing corresponding response suggestions;

[0049] S5. Display financial market trends, feedback information fluctuations, hazard assessment results and corresponding response recommendations, and generate a comprehensive report;

[0050] The method of using the emotional tendency of financial feedback information in combination with user characteristics to conduct customized feedback information hazard assessment for different users, and combining historical data to predict the impact of feedback information on financial market behavior includes the following steps:

[0051] The emotional tendency of financial feedback information is used in combination with user characteristics, trust, interest similarity, risk preference and investment strategy to conduct customized feedback information hazard assessment for different users; and the impact of feedback information on financial market behavior is predicted through time series analysis and prediction models combined with historical data.

[0052] Compared with the prior art, the present invention provides a feedback information hazard degree analysis system and method based on artificial intelligence, which has the following beneficial effects:

[0053] (1) By integrating the data collection module, text data analysis module, hazard level assessment module, early warning response module and visualization display module, it is possible to accurately analyze the sentiment tendency and user characteristics of feedback information, and provide customized hazard level assessment based on these analyses. The system's prediction capability and early warning mechanism can respond to market changes in a timely manner. At the same time, the intuitive visualization display module helps users to easily understand and analyze complex financial data and feedback information dynamics, thereby providing financial market participants with a powerful data-driven decision support tool, helping them to better understand market dynamics and make more informed decisions.

[0054] (2) By setting up a customized feedback information hazard level analysis module, it is possible to provide highly personalized and accurate analysis results. By combining the user's personal characteristics and behavioral data, it is able to provide each user with a customized feedback information hazard assessment. It not only analyzes the emotional tendency of the feedback information, but also deeply considers the user's risk preference and investment strategy, thereby ensuring the relevance and practicality of the assessment results, which helps investors make more appropriate decisions based on their own circumstances, improves the efficiency and accuracy of decision-making, and can better meet the user's usage needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0056] Figure 1 It is a structural block diagram of a feedback information hazard degree analysis system based on artificial intelligence according to an embodiment of the present invention;

[0057] Figure 2 It is a flow chart of a method for analyzing the hazard level of feedback information based on artificial intelligence according to an embodiment of the present invention.

[0058] In the figure:

[0059] 1. Data collection module; 11. Financial data acquisition module; 12. User behavior data acquisition module; 2. Text data analysis module; 21. Sentiment tendency analysis module; 22. User feature analysis module; 3. Hazard level assessment module; 31. Customized feedback information hazard level analysis module; 32. Financial market impact prediction module; 4. Early warning response module; 5. Visual display module. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] According to an embodiment of the present invention, a system and method for analyzing the hazard level of feedback information based on artificial intelligence are provided.

[0062] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to one embodiment of the present invention, a feedback information hazard degree analysis system based on artificial intelligence is provided, and the feedback information hazard degree analysis system based on artificial intelligence is used for financial feedback information. The feedback information hazard degree analysis system based on artificial intelligence includes a data acquisition module 1, a text data analysis module 2, a hazard degree assessment module 3, an early warning response module 4 and a visual display module 5;

[0063] The data collection module 1 is used to obtain real-time online financial feedback information data and user behavior data including websites, forums, and social platforms;

[0064] Financial feedback information usually refers to the public opinions and attitudes of the public, media or market participants in the financial and economic fields on specific events or topics. This feedback information can affect market dynamics, investment decisions and economic policies. The scope of financial feedback information is wide, mainly including the following categories:

[0065] 1) Company-related feedback information: including company financial report releases, executive changes, mergers and acquisitions, product launches, etc.

[0066] 2) Economic policy and regulatory feedback information: economic policies, regulatory changes, tax policies, etc. released by regulatory agencies.

[0067] 3) Market dynamic feedback information: large fluctuations or trend changes involving the stock market, bond market, foreign exchange market, commodity market, etc.

[0068] 4) Macroeconomic feedback information: involves changes in macroeconomic indicators such as GDP growth, inflation rate, and unemployment rate.

[0069] 5) Financial Innovation and Technology Feedback: Discussions on emerging technologies and trends such as FinTech and blockchain.

[0070] Specifically, the data acquisition module 1 includes a financial data acquisition module 11 and a user behavior data acquisition module 12;

[0071] The financial data acquisition module 11 is used to acquire real-time online financial feedback information data including websites, forums, and social platforms using crawler technology, and to perform correlation analysis on the real-time online financial feedback information data, and to eliminate data with correlation lower than a preset correlation threshold according to the analysis results;

[0072] In this embodiment, when judging the relevance of news, some news may include many listed companies at the same time. Although they can be regarded as hitting the keywords of listed companies, their importance is not high. Too much such news will increase the workload of push and user viewing, and they are not news that users want to pay attention to. At this time, it is necessary to automatically analyze whether the news is important by analyzing the frequency of keyword occurrence, industry relevance, and the importance of keywords, and feedback to the user.

[0073] The user behavior data acquisition module 12 is used to acquire the user's characteristic data and historical investment data using crawler technology. The characteristic data includes the user's age information, gender information and occupation information, and the historical investment data includes the user's historical transaction data and preference settings.

[0074] The text data analysis module 2 is used to analyze the acquired real-time network financial feedback information data and user behavior data to obtain the emotional tendency and user characteristics of the financial feedback information;

[0075] Specifically, the text data analysis module 2 includes a sentiment analysis module 21 and a user feature analysis module 22;

[0076] The sentiment analysis module 21 is used to identify the sentiment in the financial feedback information data through sentiment analysis technology, and the sentiment includes positive, neutral and negative;

[0077] In this embodiment, when judging whether a piece of news is negative, neutral, or positive, specific keywords such as performance decline, reduction of holdings, etc. can be provided for the investor relations management industry to adjust the algorithm and train the model, so that the judgment of sentiment analysis is more accurate and the information can be pushed to users more accurately.

[0078] Specifically, when using sentiment analysis technology to identify sentiment tendencies in financial feedback information data, the following steps are included:

[0079] Data preprocessing: Clean the data and remove useless information, such as non-text elements (pictures, video links, etc.). Standardize the text data, such as removing special characters and unifying upper and lower case.

[0080] Text analysis: Use natural language processing (NLP) technology to extract key information from the text, such as keywords, phrases, etc. Use word segmentation technology to divide the text into smaller units for easy analysis.

[0081] Sentiment Identification: Apply sentiment analysis models to identify the sentiment of text. These models are usually based on machine learning and can learn to identify sentiment in text using training data. Sentiment analysis models classify text as positive, neutral, or negative. This can be based on factors such as the sentiment of the words, context, sentence structure, etc.

[0082] Result evaluation and optimization: Evaluate the sentiment analysis results to check their accuracy and reliability. Adjust and optimize the model based on the evaluation results to improve the accuracy of sentiment recognition.

[0083] Application and feedback: Apply sentiment analysis results to areas such as feedback information monitoring and trend analysis. Collect user feedback and actual application results to further optimize the model and analysis process.

[0084] In the financial field, sentiment analysis is of great value in understanding market sentiment and predicting market trends. For example, a large amount of positive financial feedback information may indicate market optimism, while negative feedback information may indicate market concerns or uncertainty. By accurately identifying and analyzing these sentiment tendencies, we can better understand market dynamics and investor behavior.

[0085] The user characteristic analysis module 22 is used to analyze the user's characteristic data and historical investment data to obtain the user's risk preference and investment strategy.

[0086] Specifically, the risk preference and investment strategy of the user are analyzed based on the user's characteristic data and historical investment data, including the following steps:

[0087] User characteristic data collection: Collect basic information of users, such as age, occupation, income level, educational background, etc. Obtain relevant information such as user's investment experience, financial goals and investment purpose.

[0088] Historical investment data analysis: Analyze the user's historical investment records, including the asset category, investment scale, transaction frequency, etc. Study the user's investment behavior under different market conditions, such as their reaction to market fluctuations.

[0089] Risk appetite assessment: Assess the user's attitude towards risk through their investment history and behavior. For example, frequent high-risk investments may indicate that the user has a high risk appetite. Methods such as questionnaires can be used to obtain information about risk tolerance directly from users.

[0090] Investment strategy analysis: Analyze the user's investment strategy based on their historical investment decisions. For example, preference for long-term investment, diversified investment, etc. Consider the user's financial goals and investment period to determine the suitability of their investment strategy.

[0091] Comprehensive data analysis: Integrate user characteristics and historical investment data, and use data analysis and statistical methods to form a comprehensive understanding of user risk preferences and investment strategies. Advanced technologies such as machine learning can be used to improve the accuracy and efficiency of analysis.

[0092] Through these steps, we can better understand the risk preferences and investment strategies of users, so as to provide investment advice and services that better meet their needs and expectations. This analysis is of great significance to improving customer satisfaction and investment performance.

[0093] The harm level assessment module 3 is used to conduct customized feedback information harm level assessment for different users based on the emotional tendency of financial feedback information and user characteristics, and to predict the impact of feedback information on financial market behavior based on historical data;

[0094] Specifically, the hazard level assessment module 3 includes a customized feedback information hazard level analysis module 31 and a financial market impact prediction module 32;

[0095] The customized feedback information hazard level analysis module 31 is used to conduct customized feedback information hazard level assessment for different users based on the emotional tendency of financial feedback information combined with user characteristics, trust, interest similarity, risk preference and investment strategy;

[0096] The customized feedback information hazard level analysis module 31 includes a user similarity calculation module, a similar user determination module, a feedback information similarity calculation module, a similar feedback information determination module, and a hazard level determination module; it should be noted here that in this embodiment, certain parameters need to be quantified when calculating similarity.

[0097] The user similarity calculation module is used to calculate the comprehensive user similarity between the target user and other users in the historical feedback information database, and to sort the comprehensive user similarities; specifically, it includes:

[0098] The age similarity, gender similarity, occupation similarity, trust and interest similarity between the target user and other users in the historical feedback information database are calculated respectively, and the user similarity is calculated based on the age similarity, gender similarity, occupation similarity, trust and interest similarity; the calculation formula is as follows:

[0099]

[0100] D (m,n) =α·Q a (m,n)+β·Q b (m,n)+θ·Q c (m,n)

[0101]

[0102] Where, X (m,n) Indicates the similarity of user interests, D (m,n) represents the user feature similarity, T (m,n) represents the user trust, C(m) represents the feedback information set liked by user m, C(n) represents the feedback information set liked by user n, and C(i) represents the common feedback information set of user m and user n; α, β and θ represent the weights of gender feature similarity, age feature similarity and occupation feature similarity respectively, and α∈[0,1], β∈[0,1], θ∈[0,1], α+β+θ=1, Q a (m,n) represents the similarity of gender characteristics between user m and user n, Q b (m,n) represents the similarity of age characteristics between user m and user n, Q c (m,n) represents the similarity of occupational characteristics between user m and user n, P mk represents the feature value of the kth user of user m, Pnk represents the feature value of the kth user of user n, represents the average feature value of user m, represents the average feature value of user n; H represents user activity, L represents user reputation, λ represents the weight of local trust, F(m) represents the neighbors of user m, F(n) represents the neighbors of user n, j represents the number of labels annotated by the user for all feedback information, and the threshold J is less than the sum of the number of labels of the most active user, g i represents the degree of node i, g max It represents the maximum degree of a node in the trust network, and the value of L is [0,1].

[0103] The risk preference similarity and investment strategy similarity between the target user and other users in the historical feedback information database are calculated separately, and the user comprehensive similarity is obtained by combining the user similarity calculation. The user comprehensive similarity is sorted in descending order according to the similarity value.

[0104] The calculation formula of the user comprehensive similarity is:

[0105] S 综合 =ω1·S 用户 +ω2·S 偏好 +ω3·S 策略

[0106]

[0107] In the formula, S 综合 represents the comprehensive similarity of users, S 用户 represents user similarity, ω1 represents the weight of user similarity, S 偏好 represents the risk preference similarity, ω2 represents the weight of risk preference, S 策略 represents the similarity of investment strategy, which includes asset allocation, market time, risk level, investment period, etc., ω3 represents the weight of investment strategy; R m represents the risk preference of user m, R n represents the risk preference of user n. The risk preference value can be obtained through questionnaire surveys, historical investment behavior analysis, etc. It is usually a numerical value that represents the user's risk tolerance or preference. R represents the maximum possible risk preference difference, which depends on the range of risk preference scores. I represents the number of investment strategy parameters. O mi represents the i-th investment strategy parameter of user m, O ni represents the i-th investment strategy parameter of user n, and Z represents the maximum possible strategy difference, which depends on the range of each investment strategy parameter.

[0108] The similar user determination module is used to obtain the user most similar to the target user according to the ranking result of the user comprehensive similarity, and analyze and obtain the harm degree of the first initial feedback information corresponding to the most similar user;

[0109] The feedback information similarity calculation module is used to calculate the comprehensive similarity of feedback information between the target feedback information and other feedback information in the historical feedback information database, and sort the comprehensive similarity of the feedback information; specifically includes:

[0110] The sentiment tendency similarity, influence range similarity and propagation speed similarity between the target feedback information and other feedback information in the historical feedback information database are calculated respectively, and the comprehensive similarity of the feedback information is obtained based on the sentiment tendency similarity, influence range similarity and propagation speed similarity, and the comprehensive similarity of the feedback information is sorted in descending order according to the similarity value.

[0111] The calculation formula of the comprehensive similarity of the feedback information is:

[0112] Y 综合 =ε1·Y 情感 +ε2·Y 影响 +ε3·Y 传播

[0113]

[0114] Where Y 综合 represents the comprehensive similarity of feedback information, Y 情感 represents the similarity of sentiment tendency, ε1 represents the weight of sentiment tendency, Y 影响 represents the similarity of the influence range, ε2 represents the weight of the influence range, and Y 传播 represents the similarity of propagation speed, ε3 represents the weight of propagation speed; E A Indicates the emotional tendency of feedback information A, E B represents the emotional tendency of the feedback information B, E represents the maximum emotional tendency difference, which is determined according to the quantitative range of the emotional tendency, and W A represents the impact range of feedback information A, W B represents the influence range of feedback information B, W represents the maximum influence range difference, which depends on the maximum and minimum influence ranges, V A represents the propagation speed of feedback information A, V B represents the propagation speed of the feedback information B, and V represents the maximum propagation speed difference, which depends on the fastest and slowest propagation speeds.

[0115] The similar feedback information determination module is used to obtain the feedback information most similar to the target feedback information according to the sorting result of the comprehensive similarity of the feedback information, and analyze and obtain the harm degree of the second initial feedback information corresponding to the most similar feedback information;

[0116] The harm level determination module is used to calculate the final feedback information harm level based on the first initial feedback information harm level and the second initial feedback information harm level in combination with a preset weight;

[0117] The calculation formula of the final feedback information hazard level is: final feedback information hazard level = U*first initial feedback information hazard level + (1-U)*second initial feedback information hazard level, where U represents the weight of the first initial feedback information hazard level, and the value range of U is [0.1,0.4].

[0118] Specifically, in this embodiment, the harm level of financial feedback information is divided into four levels, and the corresponding relationship with the harm degree mentioned above is as follows:

[0119] Low Risk: Score range: 0-25;

[0120] Moderate Risk: Score range: 26-50;

[0121] High Risk: Score range: 51-75;

[0122] Very High Risk: Score range: 76-100;

[0123] For example, assuming that the hazard level of the first initial feedback information is 40 (medium hazard), the hazard level of the second initial feedback information is 70 (high hazard), and the weight U is set to 0.3, the final hazard level of the feedback information is calculated as: final feedback information hazard level = 0.3×40+(1-0.3)×70=61, which means that according to this calculation, the final feedback information hazard level is 61, which is a "high hazard" level.

[0124] The financial market impact prediction module 32 is used to predict the impact of feedback information on financial market behavior through time series analysis and prediction models combined with historical data.

[0125] Specifically, the impact of feedback information on financial market behavior is predicted by combining time series analysis and forecasting models with historical data, including the following steps:

[0126] Data collection and processing: Collect historical financial feedback information data, including news reports, social media content, forum discussions, etc. Collect relevant data on financial markets, such as stock prices, trading volumes, market indices, etc. Clean and pre-process the data for easy analysis.

[0127] Time Series Data Analysis: Use time series analysis techniques such as autoregressive (AR), moving average (MA), or autoregressive moving average (ARMA) to analyze trends and patterns in data over time. Identify seasonal, cyclical, or trending characteristics in the data.

[0128] Sentiment analysis and feedback information quantification: Conduct sentiment analysis on the collected feedback information data and classify it as positive, neutral or negative. Quantify the sentiment tendency to facilitate correlation analysis with market data.

[0129] Build a prediction model: Combine the results of time series analysis and sentiment analysis to build a prediction model. This includes statistical models, machine learning models, or deep learning models. The model is trained with historical data to learn the relationship between feedback information data and market behavior.

[0130] Model evaluation and optimization: Use cross-validation and model performance indicators (such as mean square error, accuracy, etc.) to evaluate the accuracy and reliability of the model. Adjust and optimize the model based on the evaluation results.

[0131] Prediction and application: Use the optimized model to predict future market behavior. Apply the prediction results to investment decision-making, risk management, market analysis and other fields.

[0132] Through these steps, it is possible to predict how feedback information affects the behavior of financial markets, providing valuable insights for investors and market analysts. For example, the model may predict that negative feedback information will cause a stock price to fall, or that positive feedback information may cause the market to rise overall. This analysis helps to better understand market dynamics and develop more effective investment strategies.

[0133] The warning response module 4 is used to monitor the financial market and feedback information dynamics in real time, trigger a warning when the degree of harm reaches a preset threshold, and provide corresponding response suggestions;

[0134] Specifically, real-time monitoring of financial markets and feedback information dynamics, and triggering warnings and providing corresponding response suggestions when the degree of harm reaches a preset threshold, include the following steps:

[0135] Establish a real-time monitoring system: Configure real-time data streams for monitoring financial markets, such as stock prices, trading volumes, market news, etc. Set up a feedback information monitoring mechanism to capture and analyze real-time feedback information data from social media, news websites, forums, etc.

[0136] Setting harm level thresholds: Based on historical data and market analysis, set different levels of harm level thresholds according to the above embodiments. These thresholds should be based on the sensitivity of the market and the risk tolerance of the enterprise. The thresholds can be fixed or dynamically adjusted to adapt to changes in the market environment.

[0137] Real-time data analysis and early warning mechanism: Real-time analysis of market data and feedback information data to monitor whether there are trends or events that reach the preset threshold. When the monitored data reaches or exceeds the preset hazard level threshold, the early warning mechanism is automatically triggered.

[0138] Early warning notifications and response suggestions: Send early warning notifications to relevant personnel or decision makers through email, mobile app push, SMS, etc. Provide response suggestions based on the current market situation and feedback information analysis. For example, it is recommended to strengthen market monitoring, adjust investment strategies, issue public relations statements, etc.

[0139] Crisis management and communication strategies: When the risk level is high, implement the crisis management plan and communication strategy prepared in advance. Provide customized response measures according to different situations, such as market intervention and communication with the public.

[0140] Follow-up tracking and analysis: Continue to track the dynamics of the market and feedback information after the early warning, and evaluate the effectiveness of the measures taken. Adjust and optimize the monitoring system and early warning mechanism to improve the efficiency and effectiveness of future responses.

[0141] Through such a system, we can effectively monitor the financial market and feedback information, timely identify potential risks, take appropriate measures to reduce the impact of risks, and protect the interests of enterprises and investors.

[0142] The visualization display module 5 is used to display financial market trends, feedback information fluctuations, hazard level assessment results and corresponding response suggestions, and generate a comprehensive report.

[0143] According to another embodiment of the present invention, Figure 2 As shown, a feedback information hazard degree analysis method based on artificial intelligence is provided, and the feedback information hazard degree analysis method based on artificial intelligence includes the following steps:

[0144] S1. Obtain real-time online financial feedback information data and user behavior data including websites, forums, and social platforms;

[0145] S2. Analyze the acquired real-time online financial feedback information data and user behavior data to obtain the emotional tendency and user characteristics of the financial feedback information;

[0146] S3. Use the sentiment tendency of financial feedback information combined with user characteristics to conduct customized feedback information harm assessment for different users, and combine historical data to predict the impact of feedback information on financial market behavior;

[0147] S4. Real-time monitoring of financial markets and feedback information dynamics, triggering warnings when the degree of harm reaches a preset threshold, and providing corresponding response suggestions;

[0148] S5. Display financial market trends, feedback information fluctuations, hazard assessment results and corresponding response recommendations, and generate a comprehensive report;

[0149] The method of using the emotional tendency of financial feedback information in combination with user characteristics to conduct customized feedback information hazard assessment for different users, and combining historical data to predict the impact of feedback information on financial market behavior includes the following steps:

[0150] The emotional tendency of financial feedback information is used in combination with user characteristics, trust, interest similarity, risk preference and investment strategy to conduct customized feedback information hazard assessment for different users; and the impact of feedback information on financial market behavior is predicted through time series analysis and prediction models combined with historical data.

[0151] To sum up, with the help of the above-mentioned technical scheme of the present invention, by integrating the data acquisition module, the text data analysis module, the hazard level assessment module, the early warning response module and the visualization display module, it is possible to accurately analyze the emotional tendency and user characteristics of the feedback information, and provide customized hazard level assessments based on these analyses. The system's predictive capabilities and early warning mechanisms can respond to market changes in a timely manner. At the same time, the intuitive visualization display module helps users to easily understand and analyze complex financial data and feedback information dynamics, thereby providing financial market participants with a powerful data-driven decision support tool, which helps them better understand market dynamics and make more informed decisions.

[0152] In addition, by setting up a customized feedback information hazard level analysis module, highly personalized and accurate analysis results can be provided. By combining the user's personal characteristics and behavioral data, it can provide each user with a customized feedback information hazard assessment. It not only analyzes the emotional tendency of the feedback information, but also deeply considers the user's risk preference and investment strategy, thereby ensuring the relevance and practicality of the evaluation results, which helps investor users make more appropriate decisions based on their own circumstances, improves the efficiency and accuracy of decision-making, and can better meet the user's usage needs.

[0153] The technical features of the above-mentioned embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. A person of ordinary skill in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program. The program can be stored in a computer-readable storage medium. When the program is executed, it includes the steps described in the above method, and the storage medium, such as ROM / RAM, a disk, an optical disk, etc.

[0154] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A feedback information hazard degree analysis system based on artificial intelligence, the feedback information hazard degree analysis system based on artificial intelligence is used for financial feedback information, characterized in that: The artificial intelligence-based feedback information hazard level analysis system comprises a data collection module (1), a text data analysis module (2), a hazard level assessment module (3), an early warning response module (4) and a visual display module (5); The data collection module (1) is used to obtain real-time online financial feedback information data and user behavior data from websites, forums, and social platforms; The text data analysis module (2) is used to analyze the acquired real-time network financial feedback information data and user behavior data to obtain the emotional tendency and user characteristics of the financial feedback information; The harm level assessment module (3) is used to conduct customized feedback information harm level assessment for different users based on the emotional tendency of the financial feedback information and user characteristics, and to predict the impact of the feedback information on financial market behavior based on historical data; The hazard level assessment module (3) includes a customized feedback information hazard level analysis module (31) and a financial market impact prediction module (32); The customized feedback information hazard level analysis module (31) is used to conduct customized feedback information hazard level assessment for different users by combining the emotional tendency of financial feedback information with user characteristics, trust, interest similarity, risk preference and investment strategy; The customized feedback information hazard level analysis module (31) comprises a user similarity calculation module, a similar user determination module, a feedback information similarity calculation module, a similar feedback information determination module, and a hazard level determination module; The user similarity calculation module is used to calculate the comprehensive user similarity between the target user and other users in the historical feedback information database, and to sort the comprehensive user similarities; The user similarity calculation module calculates the user comprehensive similarity between the target user and other users in the historical feedback information database, and sorts the user comprehensive similarities, including: Calculate the age similarity, gender similarity, occupation similarity, trust and interest similarity between the target user and other users in the historical feedback information database respectively, and calculate the user similarity based on the age similarity, gender similarity, occupation similarity, trust and interest similarity; Calculate the risk preference similarity and investment strategy similarity between the target user and other users in the historical feedback information database respectively, and calculate the user comprehensive similarity by combining the user similarity, and sort the user comprehensive similarity in descending order according to the similarity value; The similar user determination module is used to obtain the user most similar to the target user according to the ranking result of the user comprehensive similarity, and analyze and obtain the harm degree of the first initial feedback information corresponding to the most similar user; The feedback information similarity calculation module is used to calculate the comprehensive similarity of feedback information between the target feedback information and other feedback information in the historical feedback information database, and to sort the comprehensive similarity of the feedback information; The similar feedback information determination module is used to obtain the feedback information most similar to the target feedback information according to the sorting result of the comprehensive similarity of the feedback information, and analyze and obtain the harm degree of the second initial feedback information corresponding to the most similar feedback information; The harm level determination module is used to calculate the final feedback information harm level based on the first initial feedback information harm level and the second initial feedback information harm level in combination with a preset weight; The calculation formula of the final feedback information hazard level is: final feedback information hazard level = U*first initial feedback information hazard level + (1-U)*second initial feedback information hazard level, where U represents the weight of the first initial feedback information hazard level, and the value range of U is [0.1, 0.4]; The financial market impact prediction module (32) is used to predict the impact of feedback information on financial market behavior by combining time series analysis and prediction models with historical data; The early warning response module (4) is used to monitor the financial market and feedback information dynamics in real time, trigger an early warning when the degree of harm reaches a preset threshold, and provide corresponding response suggestions; The visual display module (5) is used to display financial market trends, feedback information fluctuations, hazard level assessment results and corresponding response suggestions, and generate a comprehensive report.

2. According to the artificial intelligence-based feedback information hazard analysis system of claim 1, it is characterized in that: The data collection module (1) comprises a financial data acquisition module (11) and a user behavior data acquisition module (12); The financial data acquisition module (11) is used to use crawler technology to acquire real-time online financial feedback information data including websites, forums, and social platforms, and to perform correlation analysis on the real-time online financial feedback information data, and to eliminate data with a correlation lower than a preset correlation threshold according to the analysis results; The user behavior data acquisition module (12) is used to acquire the user's characteristic data and historical investment data using crawler technology, wherein the characteristic data includes the user's age information, gender information and occupation information, and the historical investment data includes the user's historical transaction data and preference settings.

3. According to the artificial intelligence-based feedback information hazard analysis system of claim 1, it is characterized in that: The text data analysis module (2) comprises a sentiment tendency analysis module (21) and a user feature analysis module (22); The sentiment analysis module (21) is used to identify the sentiment in the financial feedback information data through sentiment analysis technology, and the sentiment includes positive, neutral and negative; The user characteristic analysis module (22) is used to analyze the user's characteristic data and historical investment data to obtain the user's risk preference and investment strategy.

4. The artificial intelligence-based feedback information hazard analysis system according to claim 1 is characterized in that: The calculation formula of the user comprehensive similarity is: S 综合 =ω1·S 用户 +ω2·S 偏好 +ω3·S 策略 In the formula, S 综合 represents the comprehensive similarity of users, S 用户 represents user similarity, ω1 represents the weight of user similarity, S 偏好 represents the risk preference similarity, ω2 represents the weight of risk preference, S 策略 represents the similarity of investment strategies, ω3 represents the weight of investment strategies; X (m,n) Indicates the similarity of user interests, D (m,n) represents the user feature similarity, T (m,n) Indicates user trust; R m represents the risk preference of user m, R n represents the risk preference of user n, R represents the maximum possible risk preference difference, I represents the number of investment strategy parameters, O mi represents the i-th investment strategy parameter of user m, O ni represents the i-th investment strategy parameter of user n, and Z represents the maximum possible strategy difference.

5. The artificial intelligence-based feedback information hazard analysis system according to claim 1 is characterized in that: The feedback information similarity calculation module calculates the comprehensive similarity of feedback information between the target feedback information and other feedback information in the historical feedback information database, and sorts the comprehensive similarity of feedback information, including: The sentiment tendency similarity, influence range similarity and propagation speed similarity between the target feedback information and other feedback information in the historical feedback information database are calculated respectively, and the comprehensive similarity of the feedback information is obtained based on the sentiment tendency similarity, influence range similarity and propagation speed similarity, and the comprehensive similarity of the feedback information is sorted in descending order according to the similarity value.

6. The artificial intelligence-based feedback information hazard analysis system according to claim 5, characterized in that: The calculation formula of the comprehensive similarity of the feedback information is: Y 综合 =ε1·Y 情感 +ε2·Y 影响 +ε3·Y 传播 Where Y 综合 represents the comprehensive similarity of feedback information, Y 情感 represents the similarity of sentiment tendency, ε1 represents the weight of sentiment tendency, Y 影响 represents the similarity of the influence range, ε2 represents the weight of the influence range, and Y 传播 represents the similarity of propagation speed, ε3 represents the weight of propagation speed; E A Indicates the emotional tendency of feedback information A, E B represents the emotional tendency of the feedback information B, E represents the maximum emotional tendency difference, which is determined according to the quantitative range of the emotional tendency, and W A represents the impact range of feedback information A, W B represents the influence range of feedback information B, W represents the maximum influence range difference, which depends on the maximum and minimum influence ranges, V A represents the propagation speed of feedback information A, V B represents the propagation speed of the feedback information B, and V represents the maximum propagation speed difference, which depends on the fastest and slowest propagation speeds.

7. A method for analyzing the hazard level of feedback information based on artificial intelligence, implemented based on the system for analyzing the hazard level of feedback information based on artificial intelligence as described in any one of claims 1 to 6, characterized in that: The artificial intelligence-based feedback information hazard degree analysis method comprises the following steps: S1. Obtain real-time online financial feedback information data and user behavior data including websites, forums, and social platforms; S2. Analyze the acquired real-time online financial feedback information data and user behavior data to obtain the emotional tendency and user characteristics of the financial feedback information; S3. Use the sentiment tendency of financial feedback information combined with user characteristics to conduct customized feedback information harm assessment for different users, and combine historical data to predict the impact of feedback information on financial market behavior; S4. Real-time monitoring of financial markets and feedback information dynamics, triggering warnings when the degree of harm reaches a preset threshold, and providing corresponding response suggestions; S5. Display financial market trends, feedback information fluctuations, hazard assessment results and corresponding response recommendations, and generate a comprehensive report; The method of using the emotional tendency of financial feedback information in combination with user characteristics to conduct customized feedback information hazard assessment for different users, and combining historical data to predict the impact of feedback information on financial market behavior includes the following steps: The emotional tendency of financial feedback information is used in combination with user characteristics, trust, interest similarity, risk preference and investment strategy to conduct customized feedback information hazard assessment for different users; and the impact of feedback information on financial market behavior is predicted through time series analysis and prediction models combined with historical data.

Images