Network public opinion risk analysis method, device and equipment based on fuzzy logic assistance, medium and product

Abstract

The invention discloses a network public opinion risk analysis method and device based on fuzzy logic assistance, equipment, a medium and a product, and relates to the field of data analysis and processing. The method comprises the following steps: acquiring public opinion data; preprocessing the public opinion data, and extracting key information from the preprocessed public opinion data based on a public opinion monitoring tool to obtain an extraction result; the key information comprises keywords and themes; based on a preset fuzzy set and a preset fuzzy logic rule, performing fuzzy reasoning processing on the extraction result to obtain a reasoning result; the preset fuzzy set is a set of corresponding membership association relationships when the extraction result is mapped to a discourse domain interval through a membership function; the preset fuzzy logic rule is determined based on a membership function; analyzing and processing the reasoning result by adopting a defuzzification method to obtain an analysis result; the analysis result comprises a risk level. The invention aims to flexibly and accurately realize public opinion analysis and improve the public opinion monitoring and analysis efficiency.

Description

Technical Field

[0001] The present application relates to the field of data analysis and processing, and in particular to a method, device, equipment, medium and product for analyzing network public opinion risk assisted by fuzzy logic. Background Art

[0002] With the rapid development of the Internet and social media, the speed and scope of information dissemination have greatly increased. The diversity and complexity of online public opinion, that is, the public's attitudes and opinions on specific events, people or issues on the Internet, make it difficult for traditional public opinion monitoring and analysis methods to cope with it.

[0003] The existing processing methods include traditional public opinion monitoring technology and modern public opinion monitoring technology. Traditional public opinion monitoring technology is achieved through keyword search, manual monitoring and statistical analysis.

[0004] Keyword search: Search and collect relevant information through preset keywords, but it is easily limited by keyword settings. Manual monitoring: Relying on manual analysis and judgment, it is inefficient and difficult to cope with massive amounts of information. Statistical analysis: Public opinion analysis based on statistical methods is difficult to handle complex and ambiguous information.

[0005] Regarding modern public opinion monitoring technology:

[0006] Natural Language Processing (NLP): uses machine learning and deep learning techniques to perform text classification, sentiment analysis, etc., but its ability to handle ambiguous and uncertain information is limited.

[0007] Social network analysis: By analyzing the information dissemination path and influence in social networks, key nodes and opinion leaders can be discovered, but the complexity is relatively high.

[0008] Therefore, it is crucial to implement public opinion analysis flexibly and accurately, and to improve the efficiency of public opinion monitoring and analysis. Summary of the invention

[0009] The purpose of this application is to provide a fuzzy logic-assisted network public opinion risk analysis method, device, equipment, medium and product, which can flexibly and accurately implement public opinion analysis and improve the efficiency of public opinion monitoring and analysis.

[0010] To achieve the above objectives, this application provides the following solutions:

[0011] In the first aspect, the present application provides a fuzzy logic-assisted network public opinion risk analysis method, comprising:

[0012] Obtaining public opinion data; the public opinion data includes: text information data, picture information data and video information data;

[0013] Preprocessing the public opinion data, and extracting key information from the preprocessed public opinion data based on a public opinion monitoring tool to obtain an extraction result; the key information includes keywords and topics;

[0014] Based on a preset fuzzy set and a preset fuzzy logic rule, the extraction result is subjected to fuzzy reasoning to obtain a reasoning result; the preset fuzzy set is a set of corresponding membership association relationships when the extraction result is mapped to a domain interval through a membership function; the preset fuzzy logic rule is determined based on the membership function;

[0015] The defuzzification method is used to analyze and process the inference result to obtain an analysis result; the analysis result includes a risk level.

[0016] Optionally, the fuzzy logic-assisted network public opinion risk analysis method further includes:

[0017] A public opinion response strategy is determined based on the analysis results; the public opinion response strategy is to prompt users to release information, guide public opinion or conduct crisis public relations.

[0018] Optionally, the public opinion data is preprocessed, and key information is extracted from the preprocessed public opinion data based on a public opinion monitoring tool to obtain an extraction result, which specifically includes:

[0019] Performing data cleaning and deduplication processing on the public opinion data to obtain public opinion processing data;

[0020] The public opinion monitoring tool is used to extract key information from the public opinion processing data based on word frequency-inverse document frequency or keyword extraction algorithm to obtain extraction results.

[0021] Optionally, based on a preset fuzzy set and a preset fuzzy logic rule, fuzzy reasoning is performed on the extraction result to obtain a reasoning result, which specifically includes:

[0022] Determine fuzzy reasoning parameters according to the extraction results; the fuzzy reasoning parameters include: sentiment score, popularity and propagation speed;

[0023] Based on a preset fuzzy set, determining the degree of membership according to the fuzzy reasoning parameters;

[0024] Based on preset fuzzy logic rules, the reasoning result is determined according to the fuzzy reasoning parameters and the membership degree.

[0025] Optionally, determining fuzzy reasoning parameters according to the extraction result specifically includes:

[0026] The sentiment analysis model is used to perform sentiment classification processing on the segmented text to obtain the sentiment score; the sentiment analysis model includes a language model BERT and a long short-term memory model LSTM; wherein the language model BERT is used to obtain the contextual relationship of the segmented text and perform sentiment analysis; the long short-term memory model LSTM is used to analyze the sequence data corresponding to the segmented text; the segmented text is obtained by segmenting the extraction result into words or phrases;

[0027] Based on the extraction results, H(t)=αN posts (t)+βN comments( t)+γN shares (t) determining the heat;

[0028] Among them, H(t) is the heat at time t; N p osts(t) is the number of posts at time t; N c omments(t) is the number of comments at time t; N shares (t) is the number of forwarding at time t; α, β, and γ are all indicator weight coefficients;

[0029] The propagation speed is determined according to the heat; indicators of the propagation speed include a growth rate and a diffusion rate.

[0030] Optionally, a defuzzification method is used to analyze and process the inference result to obtain an analysis result, which specifically includes:

[0031] use The inference result is converted to obtain a conversion result; wherein y * is the conversion result; y is the inference result; μ(y) is the maximum value of the membership function of all risk levels;

[0032] Based on the setting result mapping relationship, the analysis result is determined according to the conversion result.

[0033] In the second aspect, the present application provides a network public opinion risk analysis device based on fuzzy logic assistance, comprising:

[0034] A data acquisition module is used to acquire public opinion data; the public opinion data includes: text information data, picture information data and video information data;

[0035] An extraction module is used to pre-process the public opinion data and extract key information from the pre-processed public opinion data based on the public opinion monitoring tool to obtain an extraction result; the key information includes keywords and topics;

[0036] A processing module, for performing fuzzy reasoning on the extraction result based on a preset fuzzy set and a preset fuzzy logic rule to obtain a reasoning result; the preset fuzzy set is a set of corresponding membership association relationships when the extraction result is mapped to a domain interval through a membership function; the preset fuzzy logic rule is determined based on the membership function;

[0037] The analysis module is used to analyze and process the reasoning result by using a defuzzification method to obtain an analysis result; the analysis result includes a risk level.

[0038] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned fuzzy logic-assisted network public opinion risk analysis method.

[0039] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned fuzzy logic-assisted network public opinion risk analysis method.

[0040] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned fuzzy logic-assisted network public opinion risk analysis method.

[0041] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0042] The present application provides a method, device, equipment, medium and product for risk analysis of online public opinion assisted by fuzzy logic. Fuzzy logic, as a mathematical tool for dealing with uncertainty and ambiguity, can better solve the limitations of traditional methods in dealing with fuzzy and complex problems. Therefore, introducing fuzzy logic into online public opinion monitoring and analysis can provide a more accurate and flexible solution. By collecting public opinion data, after a series of preprocessing, features, that is, key information, are extracted, and fuzzy reasoning is performed on these features through preset fuzzy logic rules, so as to derive the risk level of public opinion, so as to formulate subsequent corresponding response strategies. As a result, the present application can flexibly and accurately implement public opinion analysis, and improve the efficiency of public opinion monitoring and analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 This is a flow chart of the network public opinion risk analysis method based on fuzzy logic assistance;

[0045] Figure 2 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0047] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0048] This application aims to process and analyze online public opinion data through fuzzy logic theory, that is, preset fuzzy logic rules, so as to assess potential risks. Specifically, this method involves automatically collecting public opinion data from various Internet platforms and social media, and converting the data into a format suitable for analysis after a series of preprocessing steps. Then, natural language processing technology or public opinion monitoring tools are used to extract features from the data, and these features are evaluated through fuzzy logic to derive the level of public opinion risk and the corresponding response strategy.

[0049] The core of this application is to use the advantages of fuzzy logic to deal with the uncertainty and ambiguity in public opinion data, and to solve the limitations of traditional public opinion analysis methods in dealing with complex and changeable online public opinion environments. This method can not only process massive amounts of data, but also update and analyze the latest public opinion information in real time, thereby providing users with timely and accurate risk assessment results. Ultimately, it can achieve better monitoring and management of online public opinion, and prevent and respond to potential risk events.

[0050] like Figure 1 As shown, the present application provides a network public opinion risk analysis method based on fuzzy logic assistance, the method comprising:

[0051] Step 100: Obtain public opinion data. Public opinion data includes: text information data, image information data and video information data.

[0052] Step 200: Preprocess the public opinion data, and extract key information from the preprocessed public opinion data based on the public opinion monitoring tool to obtain extraction results. The key information includes keywords and topics.

[0053] In one embodiment, the public opinion data is preprocessed, and key information is extracted from the preprocessed public opinion data based on a public opinion monitoring tool to obtain an extraction result, which specifically includes:

[0054] The public opinion data is cleaned and deduplicated to obtain the public opinion processing data; the public opinion monitoring tool is used to extract key information from the public opinion processing data based on the word frequency-inverse document frequency or keyword extraction algorithm to obtain the extraction results.

[0055] Step 300: Based on the preset fuzzy set and the preset fuzzy logic rule, the extraction result is subjected to fuzzy reasoning to obtain the reasoning result. The preset fuzzy set is the set of corresponding membership association relationships when the extraction result is mapped to the domain interval through the membership function; the preset fuzzy logic rule is determined based on the membership function.

[0056] In one embodiment, based on a preset fuzzy set and a preset fuzzy logic rule, fuzzy reasoning is performed on the extraction result to obtain a reasoning result, which specifically includes:

[0057] The fuzzy reasoning parameters are determined according to the extraction results; the fuzzy reasoning parameters include: sentiment score, heat and propagation speed.

[0058] Based on the preset fuzzy set, the membership degree is determined according to the fuzzy reasoning parameters.

[0059] Based on the preset fuzzy logic rules, the reasoning results are determined according to the fuzzy reasoning parameters and membership degrees.

[0060] Step 400: Use a defuzzification method to analyze and process the inference result to obtain an analysis result, which includes a risk level.

[0061] Among them, the defuzzification method is used to analyze and process the reasoning results to obtain the analysis results, which specifically include:

[0062] use The inference result is converted to obtain a conversion result; wherein y * is the conversion result; y is the inference result; μ(y) is the maximum value of the membership function of all risk levels.

[0063] Based on the set result mapping relationship, the analysis result is determined according to the conversion result.

[0064] As an optional implementation method, the network public opinion risk analysis method based on fuzzy logic assistance also includes:

[0065] Determine the public opinion response strategy based on the analysis results; the public opinion response strategy is to prompt users to release information, guide public opinion or conduct crisis public relations. In addition, it also includes displaying the public opinion response strategy.

[0066] As an optional implementation, determining the fuzzy reasoning parameters according to the extraction results specifically includes:

[0067] The sentiment analysis model is used to perform sentiment classification on the segmented text to obtain the sentiment score; the sentiment analysis model includes the language model BERT and the long short-term memory model LSTM; among them, the language model BERT is used to obtain the contextual relationship of the segmented text and perform sentiment analysis; the long short-term memory model LSTM is used to analyze the sequence data corresponding to the segmented text; the segmented text is obtained after the extraction result is segmented into words or phrases.

[0068] Based on the extraction results, H(t)=αN posts (t)+βN comments (t)+γN shares (t) determines the heat; where H(t) is the heat at time t; N posts (t) is the number of posts at time t; N comments (t) is the number of comments at time t; N shares (t) is the number of forwarding at time t; α, β, and γ are all indicator weight coefficients.

[0069] The speed of spread is determined based on the heat; indicators of the speed of spread include growth rate and diffusion rate.

[0070] In practical applications, the design ideas of the method mentioned in this application and the corresponding process steps are as follows:

[0071] Step 1: Clarify the objectives and scope of online public opinion assessment.

[0072] First, identify the online public opinion events or topics that need to be evaluated, and determine their importance, scope of impact, and potential risks. According to the evaluation object, set specific evaluation goals, such as predicting the development trend of public opinion, evaluating the level of social stability risk, etc.

[0073] Although there are no explicit regulations on the classification of online public opinion trends, we can refer to the classification principles of similar events for classification. Generally speaking, online public opinion events can be classified according to factors such as their influence, scope of dissemination, and social attention. For example, they can be divided into four levels: mild, moderate, severe, and extremely severe:

[0074] Mild public opinion situation: the general public pays very little attention to public opinion events, only a small number of social groups pay attention to public opinion events, the spread is weak, the speed is slow, and the degree of damage to society is low.

[0075] Moderate public opinion situation: Public opinion events begin to attract the attention of more social groups, and the intensity and speed of their dissemination have increased, but have not yet reached the level of widespread dissemination, and their impact on society is relatively controllable.

[0076] Severe public opinion situation: Public opinion events have aroused widespread social attention and discussion, spread quickly and widely, and have had a certain impact on social stability and development. More forceful measures need to be taken to deal with them.

[0077] Extremely severe public opinion situation: The public opinion incident is extremely serious, causing great concern and panic in the whole society, posing a major threat or causing serious impact on social stability, economic development, public interests, etc., and it is necessary to immediately activate the emergency plan for a comprehensive and thorough response.

[0078] Step 2: Data collection and preprocessing.

[0079] Crawler technology, API interfaces, social media monitoring tools and other means can be used to comprehensively collect online public opinion data related to the evaluation object, including information in various forms such as text, pictures, and videos. The collected data is preprocessed by cleaning, denoising, extracting keywords, etc. to improve data quality and analysis efficiency. The specific implementation is as follows:

[0080] 1. Determine the specific goals of online public opinion monitoring.

[0081] (1) Clarify the monitoring topic: First, it is necessary to clarify the specific monitoring topic, such as a certain brand, industry event, etc.

[0082] (2) Determine keywords: Based on the monitored topics, list relevant keywords and phrases. These keywords should be as comprehensive as possible to cover all possible public opinion information. TF-IDF is term frequency-inverse document frequency: the importance of certain words in a specific document is determined by calculating the ratio of the frequency TF of a word in a document to the prevalence IDF of the word in the entire document collection. Often used for preliminary keyword extraction. Keyword extraction algorithms such as TextRank: a graph-based ranking algorithm, similar to PageRank, used to automatically extract representative keywords from documents. This method usually does not require supervised learning and is suitable for unlabeled data sets. Topic models such as LDA: through techniques such as latent semantic analysis, identify topics in a text collection and extract representative keywords from them.

[0083] (3) Set the time range: clearly define the time range for monitoring, which can be long-term (such as the whole year) or short-term (such as during a certain activity or event).

[0084] (4) Determine the geographical scope: According to monitoring needs, set the geographical scope of public opinion monitoring, such as the national scope, specific province or city.

[0085] 2. Select and configure data crawler tools.

[0086] (1) Select crawler tools: Choose the appropriate crawler tool based on your monitoring objectives and technical capabilities. Scrapy and BeautifulSoup are commonly used Python crawler frameworks, but there are other tools such as Scrapinghub and 80legs to choose from.

[0087] (2) Write crawler rules:

[0088] Collect website URLs related to the monitoring topic to form an initial URL list. Write data extraction rules based on the web page structure, such as using XPath or CSS selectors to locate the data to be captured.

[0089] (3) Configure the crawler:

[0090] When writing a crawler, you should first check the robots.txt file of the target website to ensure that the crawler behavior complies with the website's crawling policy. Set up proxies and delays: In order to avoid excessive pressure on the target website, you can set the crawler to use a proxy server and set a reasonable request delay.

[0091] 3. Use public opinion monitoring tools.

[0092] Choose a professional public opinion monitoring tool, set monitoring keywords and topics in the tool, and use the tool's automatic classification and preliminary analysis functions.

[0093] (1) Select a public opinion monitoring tool: Based on your monitoring needs, choose a professional public opinion monitoring tool, such as Hootsuite, Brandwatch, Crimson Hexagon, etc.

[0094] (2) Set monitoring keywords and topics: In the public opinion monitoring tool, set the previously determined keywords and topics so that the tool can automatically capture and classify relevant information.

[0095] (3) Utilize automatic classification and preliminary analysis functions: Utilize the automatic classification and preliminary analysis functions of public opinion monitoring tools to perform preliminary processing on the captured information, such as sentiment analysis and trend prediction.

[0096] For sentiment analysis:

[0097] Dictionary method: Score the text based on sentiment dictionaries such as SentiWordNet. Sentiment dictionaries contain a large number of words with sentiment tendencies and their corresponding sentiment scores. This method is simple and fast, but the processing of context is relatively limited.

[0098] Deep learning models: Deep learning models based on RNN (Recurrent Neural Network), LSTM (Long Short-Term Memory), GRU (Gated Recurrent Unit), BERT (Bidirectional Encoder Representations from Transformers) and GPT (Generative Pre-trained Transformer) can better capture contextual information in text and are particularly suitable for processing complex emotional expressions.

[0099] Commonly used pre-trained models such as BERT can be fine-tuned to adapt to specific public opinion tasks. Hugging Face's Transformers model supports a variety of pre-trained sentiment analysis models (such as BERT-based sentiment classifiers), and can directly use the API to perform sentiment classification on large-scale text.

[0100] Trend Prediction:

[0101] Feature extraction: Extract important features from time series, which may include sentiment fluctuations, changes in keyword frequency, number of social media interactions, etc.

[0102] Trend prediction model: Select an appropriate prediction model to predict future trends based on historical data. For example, predict the popularity change of a keyword or the future trend of a specific emotion.

[0103] When making trend forecasts, the technical means used are as follows:

[0104] Time series analysis models: such as ARIMA (AutoRegressive Integrated Moving Average) models, are used to predict the future trend of single-variable or multi-variable time series. This method is suitable for linear and stationary data.

[0105] Regression model: Linear regression or nonlinear regression can be used to predict the trend of certain continuous indicators, such as changes in public opinion heat.

[0106] Machine Learning Algorithms: Ensemble learning algorithms such as Random Forest and XGBoost can also be used for trend prediction, especially when combining multiple features.

[0107] Deep learning models: Neural networks such as LSTM (Long Short-Term Memory) and GRU (Gated Recurrent Unit) are particularly suitable for capturing long-term dependencies and can achieve good results in public opinion trend prediction. More complex models such as Transformer can also be used to predict the trend of multivariate time series.

[0108] Natural language generation model: Some advanced language generation models (such as GPT) can be used to generate possible trends in future public opinion, especially in complex scenarios, by combining contextual information for prediction.

[0109] Automatic trend prediction tools:

[0110] Public opinion monitoring platforms: Commercial platforms such as Brandwatch and Meltwater usually have built-in trend analysis and forecasting functions, and users can generate trend reports based on social media and news data.

[0111] Google Trends: used to analyze the trend changes of keywords in Google searches, which can assist in public opinion prediction.

[0112] Custom analysis tools: Develop your own time series trend forecasting tools by combining Python’s Prophet (a trend forecasting tool developed by Facebook) or the statsmodels library.

[0113] 4. Implement real-time monitoring and dynamic analysis.

[0114] Set up system tools for real-time data collection and analysis. Use the dynamic tracking function of the tool to monitor the immediate changes in public opinion.

[0115] (1) Set up system tools: Configure public opinion monitoring tools and set the frequency and method of real-time data collection.

[0116] (2) Dynamic tracking: Utilize the dynamic tracking function of public opinion monitoring tools to monitor the immediate changes in public opinion in real time, including newly emerging keywords, changes in sentiment, etc.

[0117] 5. Manual screening and preliminary analysis.

[0118] Designate specialists to manually screen the automatically collected data to remove irrelevant or duplicate information. Integrate the data collected by the crawler with the data from the public opinion monitoring tool.

[0119] 6. Preliminary data visualization.

[0120] Use data visualization tools, such as Tableau or PowerBI, to present the analysis results in the form of charts, word clouds, etc. Visualization helps to intuitively understand the distribution, trend, and intensity of public opinion.

[0121] Step 3: Construct a fuzzy logic evaluation model for online public opinion, that is, determine the preset fuzzy logic rules.

[0122] According to the characteristics of online public opinion, a series of fuzzy sets are defined, such as public opinion heat, emotional tendency, transmission speed, etc., and their membership functions are determined. Based on expert experience, historical cases and data analysis results, a fuzzy rule base is established to describe the fuzzy relationship in public opinion risk assessment.

[0123] 1. Fuzzy sets.

[0124] In classical set theory, an element either belongs to a set or not. Fuzzy set theory relaxes this restriction, allowing elements to belong to a set with a certain "membership degree", which is between 0 (not belonging at all) and 1 (completely belonging). Let X be a non-empty set, called the domain.

[0125] A fuzzy set A on X can be expressed as: A = {x, μ A (x)|x∈X}where μ A (x) is the membership function, which maps each element in the set X, that is, the public opinion data x obtained in the previous article, to a real number in the interval [0, 1], indicating the degree to which x belongs to A.

[0126] 2. Fuzzy logic.

[0127] Fuzzy logic is an extension of classical logic that allows logical operations to be processed using membership in fuzzy sets. Operators in fuzzy logic (such as "and", "or", and "not") are also defined based on membership. A (x) represents the degree to which x belongs to A, μ B (y) indicates the degree to which y belongs to B.

[0128] Fuzzy "and" (usually using the minimum operation): min(μ A (x),μ B (y)).

[0129] Fuzzy "or" (usually using the maximum operation): max(μ A (x),μ B (y)).

[0130] Fuzzy "not" (usually using negative membership): 1-μ A (x).

[0131] 3. Obtaining fuzzy public opinion parameters of online public opinion data.

[0132] In order to process online public opinion data into the parameters required for fuzzy reasoning, the following steps are required:

[0133] (1) Divide the text into individual words or phrases.

[0134] Tokens=tokenize(cleaned_text).

[0135] Tokens: The word segmentation result of the cleaned text, representing a single word or phrase in the text.

[0136] Tokenize: The process of converting a text string into tokens.

[0137] cleaned_text (cleaned text): text data after preprocessing.

[0138] (2) Use a sentiment analysis model to classify the sentiment of the text. The sentiment analysis model used is BERT (Bidirectional Encoder Representations from Transformers): BERT is a powerful pre-trained language model that can capture the contextual relationships in sentences. By fine-tuning the BERT model, it can be adapted to specific sentiment analysis tasks.

[0139] The resulting sentiment score s ranges from [-1, 1], where -1 represents extreme negativity and 1 represents extreme positivity.

[0140] s=Sentimentanalysis(Tokens)

[0141] Sentiment analysis: Natural language processing technology in BERT, used to identify and extract emotional information from text.

[0142] Tokens: The word segmentation result of the cleaned text, representing a single word or phrase in the text.

[0143] Membership function that maps sentiment scores to fuzzy sets (positive, neutral, negative):

[0144] Negative membership function μ Negative (x):

[0145]

[0146] Neutral membership function μ Neutral (x):

[0147]

[0148] Positive membership function μ Positive (x):

[0149]

[0150] (4) Heat calculation.

[0151] The heat of public opinion can usually be calculated by measuring the amount of discussion about an event or topic within a specific period of time. Common indicators include the number of posts, the number of comments, the number of reposts, etc. It can be expressed by the following formula:

[0152] H(t)=αN posts (t)+βN comments (t)+γN shares (t)

[0153] Use Gaussian Membership Function to define low, medium and high membership functions:

[0154] Low membership function μ Low (x): centered at 0, the low heat range is around 0, and the standard deviation is Therefore, low heat is mainly concentrated in the range of x close to 0. When the heat is between 0-25, use the low heat formula.

[0155]

[0156] The membership function μ Medium (x): With 50 as the center, the medium heat range is relatively wide, with a standard deviation of Therefore, the medium heat range is larger, mainly concentrated in the range of x close to 50. Use the medium heat formula when the heat is between 25-75.

[0157]

[0158] High membership function μ High (x): With 100 as the center, the high heat range is around 100, and the standard deviation is Use the high heat formula when the temperature is above 75.

[0159]

[0160] (5) Calculation of propagation speed.

[0161] The speed of communication can be calculated by measuring the rate of change of the discussion volume. Commonly used indicators include growth rate, diffusion rate, etc. It can be expressed by the following formula:

[0162]

[0163] Among them: S(t) is the propagation speed at time t. H(t) is the heat of public opinion at time t. It is the derivative of public opinion heat with respect to time, that is, the rate of change of heat.

[0164] If the data is discrete, the derivative can be approximated using the difference formula:

[0165]

[0166] Where Δt is the time interval.

[0167] Use Bell-shaped membership function to define three membership functions: slow, medium, and fast:

[0168] Slow membership function μSlow(x): The center of the membership is at x=10, the width of the function is a=15, and the shape factor b=2.2 determines the shape of the curve. Therefore, the "slow" classification standard is that the membership is highest when the heat value is close to 10. The range of use is x between 0-25.

[0169]

[0170] Medium membership function μMedium(x): The center of the membership is at x=50, the width of the function is a=20, and the shape factor b=2.2 determines the shape of the curve. Therefore, the "medium" classification standard is that the membership is highest when the heat value is close to 50. The range of use is x between 25-70.

[0171]

[0172] Fast membership function μFast(x): The center of the membership is at x=90, the width of the function is a=15, and the shape factor b=2.2 determines the shape of the curve. Therefore, the "fast" classification standard is that the membership is highest when the heat value is close to 90. The range of use is between x 70-100.

[0173]

[0174] Step 4: Fuzzy reasoning and risk assessment.

[0175] 1. Input fuzzy variables: Convert the preprocessed public opinion data into fuzzy variables and input them into the fuzzy logic evaluation model. Convert the preprocessed public opinion data (heat, sentiment, and propagation speed) into fuzzy variables. Through the membership function, convert the actual value of each indicator into a membership degree.

[0176] For example, given a public opinion data at a time t: heat H(t), sentiment tendency S(t), propagation speed V(t). These values ​​will be mapped into fuzzy membership through the membership function. Heat,Low(H(t)) represents the corresponding membership function for the lower heat, μ Heat,Medium (H(t)) The corresponding membership function when the heat is medium, μ Heat,High (H(t)) The corresponding membership function calculation when the heat is high. Sentiment,Negative (S(t)) represents the corresponding membership function when the sentiment is negative, μ Sentiment,Neutral (S(t)) is the corresponding membership function when the sentiment tendency is neutral, μ Sentiment,Positive (S(t)) is the corresponding membership function calculation when the sentiment tendency is positive. Speed，Low (V(t)) represents the corresponding membership function for the case of slower propagation speed, μ Speed,Medium (V(t)) The corresponding membership function when the propagation speed is medium, μ Speed,High (V(t)) The corresponding membership function calculation when the propagation speed is faster.

[0177] μ Heat,Low (H(t)), μ Heat,Medium (H(t)), μ Heat,High (H(t))

[0178] μ Sentiment,Negative (S(t)), μ Sentiment,Neutral (S(t)),μ Sentiment,Positive (S(t))

[0179] μ Speed，Low (V(t)), μ Speed,Medium (V(t)), μ Speed,Hiht (V(t))

[0180] 2. Fuzzy reasoning: According to the rules in the fuzzy rule base, the input fuzzy variables are reasoned and the fuzzy assessment results of public opinion risks are calculated. Based on the pre-defined fuzzy rule base, fuzzy logic is used for reasoning. The fuzzy rule base can define the following rules:

[0181] (1) If the intensity of public opinion is low, the sentiment is positive, and the speed of spread is slow, then the public opinion risk is mild.

[0182] (2) If the public opinion heat is medium, the emotional tendency is neutral, and the transmission speed is medium, then the public opinion risk is medium.

[0183] (3) If the intensity of public opinion is high, the sentiment is negative, and the speed of spread is medium, then the public opinion risk is severe.

[0184] (4) If the intensity of public opinion is high, the sentiment is negative, and the speed of spread is fast, then the public opinion risk is extremely severe.

[0185] Each rule can be calculated using the minimum operator in fuzzy logic:

[0186] μ Risk,Light =min(μ Heat,Low ,μ Sentiment,Positive ,μ Speed，Low )

[0187] μ Risk,Medium =min(μ Heat,Medium ,μ Sentiment,Neutral ,μ Speed,Medium )

[0188] μ Risk,Heavy =min(μ Heat,High ,μ Sentiment,Negative ,μ Speed,Medium )

[0189] μ Risk,Severe =min(μ Heat,High ,μ Sentiment,Negative ,μ Speed,High )

[0190] μ Risk,Light , μ Risk,Medium , μ Risk,Heavy , μ Risk,Severe These are fuzzy membership functions for different risk levels, representing mild risk, moderate risk, severe risk, and critical risk, respectively.

[0191] μ Heat,Low , μ Heat,Medium , μ Heat,High These fuzzy set membership functions representing heat represent the membership of low heat, medium heat, and high heat, respectively. Sentiment,Positive ,μ Sentiment,Neutral , μ Sentiment,Negative These fuzzy set membership functions representing the results of sentiment analysis represent the membership of positive sentiment, neutral sentiment, and negative sentiment, respectively. Sentiment represents the result of sentiment analysis, and Positive, Neutral, and Negative represent different sentiment polarities, respectively. Speed，Low , μ Speed,Medium , μ Speed,HighThese fuzzy set membership functions representing the propagation speed represent the membership of low propagation speed, medium propagation speed and high propagation speed respectively. Speed ​​is an indicator of propagation speed, and Low, Medium and High represent different speed levels respectively. The min() function represents the "minimum value" function. In fuzzy logic, the minimum value rule is usually used to determine the output membership of a certain combination of conditions, that is, the minimum value of the membership of all input conditions is taken as the output membership. Each formula determines the membership of different risk levels by taking the minimum value of the fuzzy membership function of temperature, emotion and speed.

[0192] The membership of LightRisk is determined by the minimum value between Low Temperature, Positive Sentiment, and Low Speed. The membership of Medium Risk is determined by the minimum value between Medium Temperature, Neutral Sentiment, and Medium Speed. The membership of Heavy Risk is determined by the minimum value between High Temperature, Negative Sentiment, and Medium Speed. The membership of SevereRisk is determined by the minimum value between High Temperature, Negative Sentiment, and High Speed.

[0193] 3. Defuzzification: Convert the fuzzy evaluation results into specific values ​​or levels for easy understanding and application. Convert the fuzzy evaluation results into specific values ​​or levels. The commonly used defuzzification method is the Centroid Method.

[0194] Defuzzification is a key step in fuzzy logic systems, which is used to convert the output of fuzzy sets into a clear and quantifiable value. This is because the output of fuzzy logic systems is usually a fuzzy membership function, which represents the membership of different fuzzy sets, and these fuzzy sets are essentially not directly usable for practical applications or decision-making.

[0195] The need for actual decision-making: The output result in fuzzy logic is usually a fuzzy set (such as the membership of a variable is distributed between "high", "medium" and "low"), but in actual applications, clear numerical values ​​are required to make decisions. Defuzzification converts fuzzy output into precise numerical values, which can be easily integrated with other systems or directly used in actual operations. For example, in risk assessment, defuzzification can convert the membership of "mild risk" or "severe risk" into a specific risk score, thus providing a reference for managers.

[0196] Assume that the membership function of public opinion risk is defined as follows:

[0197] μ Light (y)=max(0,1-y)

[0198] μ Medium(y) = max(0,1-|y-1|)

[0199] μ Heavy (y) = max(0,1-|y-2|)

[0200] μ Severe (y) = max(0,1-|y-3|)

[0201] The centroid method is calculated as follows:

[0202]

[0203] μ(y)=max(μ Light (y),μ Medium (y),μ Heavy (y),μ Severe (y))

[0204] By calculating y * The value can determine the specific level of public opinion risk. Combined with the specific application scenario, the result can be mapped to the corresponding risk level:

[0205] 0≤y * <1: Mild public opinion situation; 1≤y * <2: moderate public opinion situation; 2≤y * <3: severe public opinion situation; 3≤y * ≤4: Extremely severe public opinion situation.

[0206] Step 5: Risk outcome analysis and response measures formulation.

[0207] Conduct in-depth analysis of the results of fuzzy logic assessment to identify the main sources, impact and development trends of public opinion risks. According to the assessment results, formulate corresponding public opinion response strategies and measures, including information release, public opinion guidance, crisis public relations, etc., to effectively reduce the impact of public opinion risks on social stability.

[0208] 1. Result analysis.

[0209] Time series analysis: Use time series analysis methods to predict the development trend and future changes of public opinion risks.

[0210] Trend model: Build a public opinion dissemination trend model to understand the potential development direction of public opinion through simulation and prediction.

[0211] Emergency monitoring: Monitor emergencies in real time, analyze their impact on public opinion risks, and adjust response strategies in a timely manner.

[0212] 2. Develop countermeasures.

[0213] According to the evaluation results, formulate corresponding strategies and measures to deal with public opinion, and ensure that the impact of public opinion risks on social stability is effectively reduced. The following are the details:

[0214] (1) Information release.

[0215] Transparency: Timely release of official information, transparent disclosure of the truth of events, and avoidance of rumors and misleading information.

[0216] Multi-channel release: Use official media, social platforms and other channels to release information to ensure wide coverage and rapid delivery of information.

[0217] Authoritative voice: Invite authoritative experts and institutions to interpret and comment on the information to enhance the credibility and authority of the information.

[0218] (2) Guiding public opinion.

[0219] Positive guidance: By releasing positive news and positive information, we can guide public opinion in a positive direction and alleviate the public's negative emotions.

[0220] Topic guidance: guide the public to focus on relevant positive topics, divert attention from negative events, and reduce the impact of public opinion risks.

[0221] Interactive communication: Through online and offline interactive activities, we can effectively communicate with the public, listen to their opinions and suggestions, and resolve conflicts and misunderstandings.

[0222] (3) Crisis public relations.

[0223] Mild public opinion situation: Continue to monitor the development of the incident to ensure that potential changes can be discovered in a timely manner. Appropriately disclose information about the incident to eliminate the doubts of minority groups. Guide public opinion in a positive direction by releasing positive news and the true situation of the incident.

[0224] Moderate public opinion situation: Increase monitoring frequency and pay close attention to changes in public opinion. Release detailed incident information and handling progress through official channels to reduce speculation and rumors. Actively interact with the public and the media, listen to feedback and respond in a timely manner.

[0225] Severe public opinion situation: Activate crisis emergency plan, respond and handle promptly. Disclose the truth of the incident and handling measures in detail to increase information transparency. Cooperate with mainstream media to convey information through authoritative channels and clarify misunderstandings.

[0226] Extremely severe public opinion situation: Immediately launch a comprehensive emergency plan and mobilize all relevant departments and resources to handle it. Senior leaders personally participate and issue authoritative statements to stabilize public sentiment. Release information through all possible channels, including television, radio, and the Internet, to ensure information coverage. Through expert interpretation and positive guidance, guide public opinion in a positive direction. After the incident is under control, conduct a comprehensive summary and reflection, and formulate preventive measures to prevent similar incidents from happening again.

[0227] This application aims to use fuzzy logic theory, combined with the complexity and uncertainty of online public opinion data, to build an efficient and accurate online public opinion analysis and risk assessment system. The system uses technical means such as fuzzy processing, fuzzy reasoning and fuzzy comprehensive evaluation to achieve real-time monitoring, intelligent analysis and risk assessment of online public opinion information, so as to provide strong decision-making support.

[0228] Specifically, the method mentioned in this application and the corresponding technical product are mainly composed of the following structures: data acquisition layer, data processing layer, fuzzy reasoning layer, risk assessment layer, and visualization display layer. Each structure is seamlessly connected and efficiently coordinated through data interfaces and communication protocols.

[0229] The data collection layer is responsible for collecting public opinion data in real time from various network channels (such as news websites, social media, forums, etc.). It establishes connections with data sources through protocols such as HTTP and API, and passes the collected data to the data processing layer.

[0230] The data processing layer performs preprocessing work such as cleaning, deduplication, word segmentation, and sentiment analysis on the collected raw data, and extracts key information such as sentiment tendency, topic category, etc. The data processing layer receives the data from the data collection layer, and passes the results to the fuzzy reasoning layer after processing.

[0231] The fuzzy reasoning layer performs fuzzy reasoning on the processed data according to preset fuzzy rules and fuzzy sets to identify potential risk points and trends in public opinion.

[0232] The fuzzy reasoning layer receives data from the data processing layer, uses fuzzy logic algorithms to perform reasoning analysis, and passes the results to the risk assessment layer.

[0233] The risk assessment layer conducts a quantitative assessment of public opinion risks based on the results of the fuzzy reasoning layer, combined with historical data and expert knowledge, and generates a risk assessment report.

[0234] The risk assessment layer receives the results of the fuzzy reasoning layer, combines them with other data sources (such as historical databases and expert knowledge bases) for comprehensive analysis, and passes the assessment results to the visualization display layer and the user interaction layer.

[0235] The visual display layer intuitively displays the risk assessment results in the form of charts, word clouds, dashboards, etc., making it easier for users to quickly understand public opinion trends and risk conditions.

[0236] The visualization layer receives data from the risk assessment layer and displays it to users through a graphical interface, while also providing data export and sharing functions.

[0237] As the data source of the system, the data collection layer provides basic data support for subsequent data processing and analysis. The data processing layer preprocesses the raw data, extracts key information, and provides high-quality data input for subsequent fuzzy reasoning and risk assessment. The fuzzy reasoning layer uses fuzzy logic algorithms to perform intelligent reasoning on the processed data, identify potential risk points and trends in public opinion, and provide strong support for risk assessment. The risk assessment layer conducts quantitative risk assessment based on the results of the fuzzy reasoning layer and other data sources, and generates a risk assessment report with reference value. The visualization layer presents complex data analysis results to users in an intuitive and easy-to-understand form, improving the efficiency and accuracy of information transmission.

[0238] In an exemplary embodiment, a fuzzy logic-assisted network public opinion risk analysis device is provided, comprising:

[0239] A data acquisition module is used to acquire public opinion data; the public opinion data includes: text information data, picture information data and video information data;

[0240] An extraction module is used to pre-process the public opinion data and extract key information from the pre-processed public opinion data based on the public opinion monitoring tool to obtain an extraction result; the key information includes keywords and topics;

[0241] A processing module, for performing fuzzy reasoning on the extraction result based on a preset fuzzy set and a preset fuzzy logic rule to obtain a reasoning result; the preset fuzzy set is a set of corresponding membership association relationships when the extraction result is mapped to a domain interval through a membership function; the preset fuzzy logic rule is determined based on the membership function;

[0242] The analysis module is used to analyze and process the reasoning result by using a defuzzification method to obtain an analysis result; the analysis result includes a risk level.

[0243] Fuzzy logic can effectively handle uncertainty and ambiguity, which is impossible with traditional binary logic or explicit classification methods. Public opinion information is often ambiguous, and emotional expressions are not always clearly positive or negative, but exist in the middle. Fuzzy logic allows for processing and reasoning of incomplete matches, making the analysis results closer to the actual situation.

[0244] In addition, the method mentioned in this application has the advantages of flexible rule setting and adjustment.

[0245] (1) High flexibility: Through fuzzy logic rules, rules can be set and adjusted according to specific needs. For example, different fuzzy rules can be set according to different public opinion characteristics to flexibly respond to various complex situations.

[0246] (2) Fuzzy reasoning allows users to set a set of "if-then" rules, which can be modified and expanded according to actual needs to enhance the adaptability of the system.

[0247] Fuzzy logic can take into account a variety of influencing factors (such as sentiment, popularity, and speed of spread) to provide a comprehensive risk assessment. This multi-factor comprehensive analysis helps to more accurately identify public opinion risks. It can combine the fuzzy membership of multiple input variables for comprehensive judgment and reasoning, thereby obtaining a more comprehensive assessment result.

[0248] Benefits of this application:

[0249] (1) Real-time monitoring and early warning: It can realize real-time monitoring of online public opinion, timely discover and warn of negative public opinion, and help users take quick response measures to avoid the escalation of public opinion.

[0250] (2) Accurate sentiment analysis: Through fuzzy logic and natural language processing technology, it is possible to conduct accurate sentiment analysis on public opinion data, identify subtle changes in public sentiment, and provide a basis for decision-making.

[0251] (3) Comprehensive risk assessment: It can comprehensively assess public opinion risks by taking into account multiple factors such as emotion, popularity, and speed of dissemination, helping users to accurately judge the level of public opinion risk and formulate targeted response strategies.

[0252] (4) Automation and intelligence: Adopt automation and intelligent technologies to reduce manual intervention, improve the efficiency and accuracy of public opinion monitoring and analysis, and reduce operating costs.

[0253] The solution mentioned in this application can help enterprises monitor and analyze brand reputation in real time, respond to negative public opinions in a timely manner, and protect corporate image and market competitiveness. It can help governments and public institutions understand the public's attitude and reaction to policies and events, improve the transparency and credibility of public decision-making, and resolve social conflicts in a timely manner. It can help individual users protect their privacy and reputation and avoid unnecessary impacts from negative public opinions.

[0254] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 2As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a network public opinion risk analysis method based on fuzzy logic assistance is implemented.

[0255] Those skilled in the art will understand that Figure 2 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0256] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0257] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0258] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0259] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0260] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0261] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0262] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above 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.

[0263] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A fuzzy logic-assisted network public opinion risk analysis method, characterized in that: The network public opinion risk analysis method based on fuzzy logic assistance includes: Obtaining public opinion data; the public opinion data includes: text information data, picture information data and video information data; Preprocessing the public opinion data, and extracting key information from the preprocessed public opinion data based on a public opinion monitoring tool to obtain an extraction result; the key information includes keywords and topics; Based on a preset fuzzy set and a preset fuzzy logic rule, the extraction result is subjected to fuzzy reasoning to obtain a reasoning result; the preset fuzzy set is a set of corresponding membership association relationships when the extraction result is mapped to a domain interval through a membership function; the preset fuzzy logic rule is determined based on the membership function; The defuzzification method is used to analyze and process the inference result to obtain an analysis result; the analysis result includes a risk level.

2. The fuzzy logic-assisted network public opinion risk analysis method according to claim 1 is characterized in that: The network public opinion risk analysis method based on fuzzy logic assistance also includes: A public opinion response strategy is determined based on the analysis results; the public opinion response strategy is to prompt users to release information, guide public opinion or conduct crisis public relations.

3. The fuzzy logic-assisted network public opinion risk analysis method according to claim 1 is characterized in that: Preprocess the public opinion data, and extract key information from the preprocessed public opinion data based on the public opinion monitoring tool to obtain extraction results, specifically including: Performing data cleaning and deduplication processing on the public opinion data to obtain public opinion processing data; The public opinion monitoring tool is used to extract key information from the public opinion processing data based on word frequency-inverse document frequency or keyword extraction algorithm to obtain extraction results.

4. The fuzzy logic-assisted network public opinion risk analysis method according to claim 1 is characterized in that: Based on the preset fuzzy set and the preset fuzzy logic rules, the extraction result is subjected to fuzzy reasoning processing to obtain the reasoning result, which specifically includes: Determine fuzzy reasoning parameters according to the extraction results; the fuzzy reasoning parameters include: sentiment score, popularity and propagation speed; Based on a preset fuzzy set, determining the degree of membership according to the fuzzy reasoning parameters; Based on preset fuzzy logic rules, the reasoning result is determined according to the fuzzy reasoning parameters and the membership degree.

5. The fuzzy logic-assisted network public opinion risk analysis method according to claim 4 is characterized in that: Determining fuzzy reasoning parameters according to the extraction results specifically includes: The sentiment analysis model is used to perform sentiment classification processing on the segmented text to obtain the sentiment score; the sentiment analysis model includes a language model BERT and a long short-term memory model LSTM; wherein the language model BERT is used to obtain the contextual relationship of the segmented text and perform sentiment analysis; the long short-term memory model LSTM is used to analyze the sequence data corresponding to the segmented text; the segmented text is obtained by segmenting the extraction result into words or phrases; Based on the extraction results, H(t)=αN posts (t)+βN comments (t)+γN shares (t) determining the heat; Among them, H(t) is the heat at time t; N posts (t) is the number of posts at time t; N comments (t) is the number of comments at time t; N shares (t) is the number of forwarding at time t; α, β, and γ are all indicator weight coefficients; The propagation speed is determined according to the heat; indicators of the propagation speed include a growth rate and a diffusion rate.

6. The fuzzy logic-assisted network public opinion risk analysis method according to claim 1 is characterized in that: The defuzzification method is used to analyze and process the inference results to obtain analysis results, which specifically include: use The inference result is converted to obtain a conversion result; wherein y * is the conversion result; y is the inference result; μ(y) is the maximum value of the membership function of all risk levels; Based on the setting result mapping relationship, the analysis result is determined according to the conversion result.

7. A network public opinion risk analysis device based on fuzzy logic, characterized in that: The network public opinion risk analysis device based on fuzzy logic assistance includes: A data acquisition module is used to acquire public opinion data; the public opinion data includes: text information data, picture information data and video information data; An extraction module is used to pre-process the public opinion data and extract key information from the pre-processed public opinion data based on the public opinion monitoring tool to obtain an extraction result; the key information includes keywords and topics; A processing module, for performing fuzzy reasoning on the extraction result based on a preset fuzzy set and a preset fuzzy logic rule to obtain a reasoning result; the preset fuzzy set is a set of corresponding membership association relationships when the extraction result is mapped to a domain interval through a membership function; the preset fuzzy logic rule is determined based on the membership function; The analysis module is used to analyze and process the reasoning result by using a defuzzification method to obtain an analysis result; the analysis result includes a risk level.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the fuzzy logic-assisted network public opinion risk analysis method described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the fuzzy logic-assisted network public opinion risk analysis method described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the fuzzy logic-assisted network public opinion risk analysis method described in any one of claims 1 to 6 is implemented.

Images