Public opinion management method based on evidence reasoning, product, medium and equipment

Abstract

The invention discloses a public opinion management method based on evidence reasoning, a product, a medium and equipment, and relates to the technical field of data mining and analysis, and the method comprises the steps: obtaining original public opinion data of a plurality of data sources; preprocessing the original public opinion data to obtain preprocessed public opinion data; performing emotional tendency scoring on the preprocessed public opinion data by adopting an emotional analysis model and a theme generation model to obtain an emotional score; based on the emotion score, a Dempster-Shafer evidence theory is adopted to obtain a comprehensive emotion score; and performing public opinion state analysis and response strategy formulation according to the comprehensive emotion score to complete public opinion management. The public opinion management efficiency and accuracy can be effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of data mining and analysis, and in particular to a method, product, medium and equipment for public opinion management based on evidence reasoning. Background Art

[0002] In the current information age, with the rapid development of the Internet and social media, public opinion monitoring and analysis has gradually become an important demand of today's society. Timely and accurate monitoring and analysis of public opinion information is crucial. However, most traditional public opinion analysis systems rely on manual data collection and analysis, which is inefficient and prone to missing important information. With the explosive growth of data volume and the increasing complexity of public opinion situations, traditional public opinion analysis methods can no longer meet actual needs.

[0003] In recent years, automated public opinion analysis systems based on big data and artificial intelligence technologies have gradually emerged. However, the public opinion analysis systems currently on the market usually rely on a single data source, such as Weibo, WeChat or a specific news website, lack cross-platform data integration capabilities, and these public opinion analysis systems have technical bottlenecks in data capture and processing, making it difficult to capture and analyze large-scale data in real time. In addition, the sentiment analysis and topic recognition technologies used in existing public opinion analysis systems mainly rely on rules and traditional machine learning methods, with limited accuracy, and are difficult to cope with complex and changing public opinion environments. In summary, the following technical problems exist in existing public opinion analysis systems: First, incomplete data collection: existing public opinion analysis systems often rely on a single data source and lack the ability to integrate multi-platform and multi-type data, resulting in incomplete public opinion information. Second, low data processing efficiency: Due to the large amount of data, existing public opinion analysis systems take a long time to process data and are difficult to achieve real-time analysis. Third, the accuracy of sentiment analysis and topic recognition is limited: the traditional sentiment analysis and topic recognition technologies used by existing public opinion analysis systems are not accurate enough when processing complex public opinion data, and it is difficult to effectively capture the sentiment tendency and topic distribution in the data. Fourth, the evidence synthesis and conflict resolution are not perfect: the existing public opinion analysis system lacks effective evidence synthesis and conflict resolution mechanisms when processing multi-source information, resulting in unreliable analysis results. As a result, the existing public opinion analysis system has the disadvantages of low efficiency and poor accuracy in the analysis and processing of public opinion. Summary of the invention

[0004] The purpose of this application is to provide a public opinion management method, product, medium and equipment based on evidence reasoning, which can effectively improve the efficiency and accuracy of public opinion management.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In the first aspect, the present application provides a public opinion management method based on evidence reasoning, including:

[0007] Obtain raw public opinion data from multiple data sources;

[0008] Preprocessing the original public opinion data to obtain preprocessed public opinion data;

[0009] Use sentiment analysis model and topic generation model to score the sentiment tendency of preprocessed public opinion data and obtain sentiment scores;

[0010] Based on the sentiment score, the Dempster-Shafer evidence theory is used to obtain the comprehensive sentiment score;

[0011] Analyze public opinion status and formulate response strategies based on comprehensive sentiment scores to complete public opinion management.

[0012] Optionally, the obtaining of original public opinion data from multiple data sources specifically includes:

[0013] Use the official API or data crawling tools of each data source to continuously obtain original public opinion data from multiple data sources at preset intervals, and record the timestamp and data ID of the captured original public opinion data.

[0014] Optionally, preprocessing the original public opinion data to obtain preprocessed public opinion data specifically includes:

[0015] The original public opinion data is cleaned, deduplicated, standardized, segmented, noise filtered, stop word removed, and stemmed or morphologically restored to obtain the preprocessed public opinion data.

[0016] Optionally, the sentiment analysis model and the topic generation model are used to score the sentiment tendency of the preprocessed public opinion data to obtain the sentiment score, which specifically includes:

[0017] Use the sentiment analysis model to score the sentiment tendency of the preprocessed public opinion data and obtain the initial sentiment score Among them, s pos 、s neg and neu They represent the initial positive, negative, and neutral sentiment scores of the preprocessed public opinion data respectively;

[0018] Use the topic generation model to classify the preprocessed public opinion data into topics and obtain topic scores Among them, l i represents the probability of the ith topic of the preprocessed public opinion data, and k is the total number of topics;

[0019] The initial sentiment score is modified based on the topic score to obtain the sentiment score.

[0020] Optionally, the step of correcting the initial sentiment score based on the topic score to obtain the sentiment score specifically includes:

[0021] Based on topic score Use the formula and For pos 、s neg and neu Correction is performed to obtain the corrected positive, negative and neutral sentiment scores s' pos 、s' neg and s' neu ; Among them, α i is the positive sentiment weight of the ith topic; β i is the negative sentiment weight of the ith topic; i is the neutral sentiment weight of the ith topic;

[0022] According to the corrected positive, negative and neutral sentiment scores s' pos 、s' neg and s' neu Get sentiment score

[0023] Optionally, the sentiment score is based on the Dempster-Shafer evidence theory to obtain a comprehensive sentiment score, specifically including:

[0024] Based on the sentiment score of the nth preprocessed public opinion data Using formula m n ({positive})=s' pos,n 、m n ({negative})=s' neg,n and m n ({neutral})=s' neu,n The corrected positive, negative and neutral sentiment scores s' for the nth preprocessed public opinion data pos,n 、s' neg,n and s' neu,n Mapping is performed to obtain the sentiment score m of the nth pre-processed public opinion data after mapping n =(m n ({positive}),m n ({negative}),m n ({neutral})); where m n ({positive}), m n ({negative}) and m n({neutral}) are the positive, negative, and neutral sentiment scores of the nth preprocessed public opinion data after mapping;

[0025] Based on m n =(m n ({positive}),m n ({negative}),m n ({neutral})), and respectively adopt the formulas through Dempster-Shafer evidence theory and The comprehensive positive emotion score m({positive}), the comprehensive negative emotion score m({negative}) and the comprehensive neutral emotion score m({neutral}) are calculated; where K is the conflict coefficient; n=1, 2;

[0026] Based on m({positive}), m({negative}) and m({neutral}), we get the comprehensive sentiment score m=(m({positive}),m({negative}),m({neutral})).

[0027] Optionally, the public opinion status analysis and response strategy formulation based on the comprehensive sentiment score to complete public opinion management specifically includes:

[0028] When the comprehensive positive sentiment score m({positive}) is greater than or equal to 0.7, the public opinion status is generally positive; when the comprehensive positive sentiment score m({positive}) is greater than or equal to 0.4 and less than 0.7, the public opinion status is relatively positive; when the comprehensive positive sentiment score m({positive}) is less than 0.4, the public opinion status is general;

[0029] When the comprehensive negative sentiment score m({negative}) is greater than or equal to 0.7, the public opinion status is generally negative; when the comprehensive negative sentiment score m({positive}) is greater than or equal to 0.4 and less than 0.7, the public opinion status is relatively negative; when the comprehensive negative sentiment score m({positive}) is less than 0.4, the public opinion status is general;

[0030] When the comprehensive neutral sentiment score m({neutral}) is greater than or equal to 0.5, the public opinion state is neutral; when the comprehensive neutral sentiment score m({neutral}) is less than 0.5, the public opinion state is non-neutral;

[0031] Formulate response strategies based on different combinations of public opinion status to complete public opinion management.

[0032] In a second aspect, the present application provides a computer program product, including a computer program, which implements the public opinion management method based on evidence reasoning when executed by a processor.

[0033] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the public opinion management method based on evidence reasoning.

[0034] In a fourth 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 public opinion management method based on evidence reasoning.

[0035] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0036] The present application provides a public opinion management method, product, medium and equipment based on evidence reasoning. The method combines complex algorithms with logical processing methods. After obtaining the original public opinion data from multiple data sources, the original public opinion data is pre-processed, and the sentiment analysis model and the topic generation model are used to score the sentiment tendency of the pre-processed public opinion data to obtain the sentiment score. After that, the Dempster-Shafer evidence theory is used to obtain the data processing process of the comprehensive sentiment score. The method integrates advanced data processing and intelligent analysis technologies, and aims to efficiently and accurately identify the dynamics of social public opinion, and automatically or manually generate corresponding response strategies, thereby improving the efficiency and accuracy of public opinion management. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] 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.

[0038] Figure 1 A flowchart of the public opinion management method based on evidence reasoning provided for this application;

[0039] Figure 2 A schematic diagram of the structure of the public opinion management device based on evidence reasoning provided in this application. DETAILED DESCRIPTION

[0040] 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.

[0041] 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 in conjunction with the accompanying drawings and specific implementation methods.

[0042] With the rapid development of big data technology, natural language processing (NLP) technology, and artificial intelligence (AI) technology, it has become possible to develop efficient and accurate public opinion analysis methods based on these new technologies. Therefore, the research and development of the public opinion management method based on evidence reasoning provided in this application is based on the following premises: an efficient data collection system that can integrate multiple data sources across platforms is required; a preprocessing system that can quickly and accurately process large-scale data is required; the latest technology is required for high-precision sentiment analysis and topic identification; an evidence synthesis mechanism that can effectively resolve information conflicts is required. The specific process of solving these premises by the public opinion management method based on evidence reasoning provided in this application is as follows.

[0043] like Figure 1 As shown, the present application discloses a method for public opinion management based on evidence reasoning, including:

[0044] Step 1: Obtain raw public opinion data from multiple data sources.

[0045] First, the data source needs to be determined. Collect raw public opinion data from multiple data sources (such as social media, news websites, forums, etc.). Specifically, based on the target audience and industry characteristics of public opinion analysis, select representative social media platforms (such as Weibo, WeChat, Tik Tok), news websites, forums, blogs, and government announcements as data sources.

[0046] The target audience usually refers to the stakeholders served by public opinion analysis, including:

[0047] Government departments: Need to monitor public opinion in a timely manner, prevent or respond to emergencies and social conflicts, maintain social stability, and adjust policies through public opinion feedback.

[0048] Public relations and media managers: They need to understand media coverage of specific events in order to develop effective public relations strategies to enhance brand image or deal with negative public opinion.

[0049] Academic researchers: provide theoretical support by analyzing public opinion data to study social phenomena and the mechanisms of public opinion formation.

[0050] Different industries have different needs and focuses on public opinion analysis, such as:

[0051] Public service industries (such as transportation, medical care, education, etc.): Pay attention to the public's feedback on service quality and policy changes. Government websites, complaint platforms, social media, etc. are the main sources of data.

[0052] Technology industry: Pay attention to user feedback on new technologies and new products, especially discussions on product features and privacy issues. Technology news, industry blogs, social media, and technology forums (such as GitHub and Reddit) are the main data sources.

[0053] Entertainment industry: Focus on the public reaction to film and television works, celebrities, and entertainment events. Entertainment news websites, social media (such as Tik Tok, Instagram, Weibo), fan forums, review websites, etc. are the main data sources.

[0054] Government and public affairs: Focus on the public’s response to policies, discussions on social issues, and the direction of public opinion on emergencies. The main data sources are government announcements, news websites, social media, forums, etc.

[0055] After determining the data source, the official API application programming interface of each data source can be used to perform legal and compliant data crawling. If the data source does not have an open API, legal and compliant data crawling can be performed through data crawling tools. As a specific embodiment, the present application can design and maintain a customized web crawler program to ensure efficient and stable data crawling.

[0056] When performing data crawling, the public opinion management method based on evidence reasoning of this application also sets a crawling strategy: set a fixed time interval (such as every hour, every day) as the preset time, and automatically perform crawling tasks from multiple data sources at preset intervals to capture the latest public opinion trends and obtain original public opinion data. The public opinion management method based on evidence reasoning of this application monitors public opinion events that require immediate response in real time, and can use technologies such as WebSocket and StreamAPI to realize instant data capture. In addition, the timestamp of each crawl and the ID of the acquired data are recorded to ensure that only new or updated data is obtained during the next crawl to avoid repeated crawling.

[0057] Suppose the data source set is S = {S1, S2, ..., S h}, where S h is the hth data source, h is the total number of data sources. The data provided by the jth data source is D j , then the collected original public opinion dataset is

[0058] Step 2: Preprocess the original public opinion data to obtain the preprocessed public opinion data.

[0059] This application preprocesses the collected original public opinion data respectively by cleaning, deduplication, standardization, word segmentation, noise filtering, stop word removal, and stemming or lemmatization to obtain the preprocessed public opinion data, so as to improve the data quality. The specific preprocessing operations are as follows.

[0060] Deduplication: Remove duplicate data records through a hash algorithm or database query.

[0061] Standardization: Unify the data formats, such as dates, times, user IDs, text contents, etc., for subsequent data processing.

[0062] Word segmentation: Use natural language processing tools to segment the text for subsequent sentiment analysis, keyword extraction, etc.

[0063] Noise filtering: Remove noise data such as advertisements, irrelevant links, non-text contents (such as picture and video links).

[0064] Stop word removal: Remove common words that are meaningless for analysis (such as "de", "le", "zai", etc.).

[0065] Stemming or lemmatization: Restore words to their basic forms for unified processing of words in different tenses and voices.

[0066] Let the preprocessing function be P(x), and apply it to the original public opinion data set D to obtain the preprocessed public opinion data set D' = P(D). Suppose x f is a piece of original public opinion data, then the preprocessed public opinion data P(x f ) includes the word-segmented preprocessed public opinion data T(x f ) and the noise-filtered preprocessed public opinion data N(x f ), that is, P(x f ) = N(T(x f )).

[0067] In summary, efficient data collection and preprocessing are the innovations of the public opinion management method based on evidence reasoning in this application. The public opinion management method based on evidence reasoning in this application can efficiently and stably collect original public opinion data from multiple data sources (such as social media, news websites, forums, etc.), and through automated and real-time scraping strategies, ensure the timeliness and comprehensiveness of the original public opinion data. The public opinion management method based on evidence reasoning in this application can efficiently clean, deduplicate, and standardize the collected original public opinion data, improve the data quality, and provide a reliable data basis for subsequent analysis.

[0068] The data collection strategy in step 1 includes using the official API and customized web crawler programs, as well as setting preset time intervals and real-time monitoring. Through the combination of API and web crawler, the public opinion management method based on evidence reasoning in this application can efficiently capture large amounts of data under the premise of legality and compliance. The setting of fixed time intervals and real-time monitoring ensures the latest and completeness of the original public opinion data, and avoids the problems of data omissions and duplicate capture. The data preprocessing in step 2 includes operations such as deduplication, standardization, noise filtering, word segmentation, stop word removal, stem extraction or word form restoration. Through a series of preprocessing operations, the public opinion management method based on evidence reasoning in this application can effectively remove noise data, duplicate data, unify data formats, and ensure standardized processing of data, thereby improving the accuracy and reliability of data.

[0069] Step 3: Use the sentiment analysis model and topic generation model to score the sentiment tendency of the preprocessed public opinion data to obtain the sentiment score.

[0070] Step 3 of this application is evidence collection and presentation.

[0071] Each piece of preprocessed public opinion data is regarded as a piece of evidence, and the sentiment tendency is scored to obtain the initial sentiment score and assign it a basic probability assignment (BPA).

[0072] Specifically, use sentiment analysis models (such as BERT, DistilBERT, etc.) to score the sentiment tendency of each preprocessed public opinion data to obtain the positive, negative and neutral sentiment scores of the preprocessed public opinion data. Suppose the sentiment analysis model is S(x'), and its output sentiment score vector (initial sentiment score) is Among them, s pos 、s neg and neu They represent the initial positive, negative, and neutral sentiment scores of the preprocessed public opinion data, respectively.

[0073] Then, we use a topic generation model, such as a document topic generation model (Latent Dirichlet Allocation, LDA), to classify each pre-processed public opinion data into topics and obtain the probability distribution of different topics. Let the topic generation model be L(x'), and its output topic score vector (topic score) be Among them, l i represents the probability of the i-th topic of the preprocessed public opinion data, and k is the total number of topics.

[0074] According to the set emotional importance weight of each topic, the initial emotional score is weighted and adjusted based on different topics. Let the positive emotional weight, negative emotional weight and neutral emotional weight of the i-th topic be α i , β i and γ i The initial sentiment score is corrected by combining the topic score and the topic feature weight (sentiment weight) to obtain the corrected positive, negative and neutral sentiment scores s' pos 、s' neg and s' neu .

[0075]

[0076] According to the characteristics of each topic, the emotional importance weight is set: (1) The weight of formal topics such as national affairs may be higher, such as: the positive emotional weight of formal topics α formal =1.2, negative sentiment weight of formal topic β formal =1.5, the neutral sentiment weight of the formal topic γ formal =1.0. (2) The weight of common topics may be lower, such as the positive sentiment weight of common topics α casual =1.0, negative sentiment weight of common topics β casual =1.0, neutral sentiment weight of common topics γ casual =1.0.

[0077] According to the corrected positive, negative and neutral sentiment scores s' pos 、s' neg and s' neu Get sentiment score

[0078] In summary, accurate sentiment analysis and topic identification methods are innovations of the public opinion management method based on evidence reasoning in this application. The public opinion management method based on evidence reasoning in this application utilizes sentiment analysis models and topic generation models, which can accurately classify the sentiment tendencies and topics of public opinion data, and correct the sentiment scores according to the characteristics of the topics. Through the evidence collection and representation, topic identification and sentiment score correction in step 3, combined with the application of sentiment analysis models and subject generation models, it is possible to carefully analyze the sentiment tendencies and topic distribution of public opinion data, and perform weighted adjustments to sentiment scores according to topic characteristics, thereby improving the accuracy of sentiment analysis.

[0079] Step 4: Based on the sentiment score, the Dempster-Shafer evidence theory is used to obtain the comprehensive sentiment score.

[0080] Step 4 of this application is evidence synthesis and conflict resolution.

[0081] Suppose the evidence set based on the preprocessed public opinion data is E = {e1, e2, …, e g ,…,e p}, the g-th piece of evidence e g The BPA is m g (A), p is the total number of evidence. A is a proposition. According to the Dempster-Shafer theory, BPA satisfies 0≤m g (A)≤1 and where Θ is the set of all possible propositions.

[0082] In online public opinion analysis, public opinion data from different sources may conflict and be inconsistent. In order to resolve these conflicts, it is necessary to synthesize evidence from different sources. The Dempster-Shafer evidence theory provides an effective tool to synthesize multiple pieces of evidence into a comprehensive piece of evidence through the Dempster synthesis rule and calculate the comprehensive sentiment score.

[0083] Assume that the sentiment scores of two pieces of evidence e1 and e2 are and Each sentiment score contains three probabilities: positive, negative, and neutral. but For two sentiment scores and The corresponding BPAs are m1(B) and m2(C), and the calculation formula for the comprehensive sentiment score m(A) is:

[0084]

[0085] Among them, the conflict coefficient K is defined as:

[0086]

[0087] Among them, A, B, and C are all propositions.

[0088] Map the sentiment score to BPA. Based on the sentiment score of the nth preprocessed public opinion data Using formula m n ({positive})=s' pos,n 、m n ({negative})=s' neg,n and m n ({neutral})=s' neu,n The corrected positive, negative and neutral sentiment scores s' for the nth preprocessed public opinion data pos,n 、s'neg,n and s' neu,n Mapping is performed to obtain the sentiment score m of the nth pre-processed public opinion data after mapping n =(m n ({positive}),m n ({negative}),m n ({neutral})). Among them, m n ({positive}), m n ({negative}) and m n ({neutral}) are respectively the positive, negative and neutral sentiment scores of the nth pre-processed public opinion data after mapping. As a specific embodiment, assuming that the BPAs of two pieces of evidence e1 and e2 are m1 and m2 respectively, the comprehensive positive sentiment score m({positive}), the comprehensive negative sentiment score m({negative}) and the comprehensive neutral sentiment score m({neutral}) are obtained by the following formula:

[0089]

[0090] Based on m({positive}), m({negative}) and m({neutral}), we get the comprehensive sentiment score m=(m({positive}),m({negative}),m({neutral})).

[0091] In summary, effective evidence synthesis and conflict resolution methods are the innovations of the public opinion management method based on evidence reasoning of this application. The public opinion management method based on evidence reasoning of this application adopts the Dempster-Shafer evidence theory, and synthesizes evidence from multiple sources into a comprehensive evidence through the Dempster synthesis rule to solve the problems of information conflicts and inconsistencies. The evidence synthesis and conflict resolution in step 4 include the application of the Dempster-Shafer evidence theory and the Dempster synthesis rule. Through the Dempster-Shafer evidence theory, evidence from different sources can be effectively processed, conflicts and inconsistencies between evidence can be resolved, and the credibility and reliability of the comprehensive evidence can be improved.

[0092] Step 5: Analyze the public opinion status and formulate response strategies based on the comprehensive sentiment score to complete public opinion management.

[0093] This application evaluates the public opinion status through a comprehensive sentiment score m = (m({positive}), m({negative}), m({neutral})).

[0094] Positive Sentiment: m({positive})≥0.7: public opinion is generally positive; 0.4≤m({positive})<0.7: public opinion is relatively positive; m({positive})<0.4: public opinion is average.

[0095] Negative Sentiment: m({negative})≥0.7: public opinion is generally negative; 0.4≤m({negative})<0.7: public opinion is relatively negative; m({negative})<0.4: public opinion is average.

[0096] Neutral Sentiment: m({neutral})≥0.5: public opinion is neutral, m({neutral})<0.5: public opinion is not neutral.

[0097] According to different combinations of comprehensive sentiment scores, that is, different combinations of public opinion states, different response strategies are adopted to complete public opinion management. The following is an example response strategy framework, as shown in Table 1.

[0098] (1) Emotion combination 1: high positive emotion, low negative emotion

[0099] Condition: m({positive})≥0.7 and m({negative})<0.3.

[0100] Response strategy: Maintain existing policies, continue to promote positive publicity, focus on a few negative public opinion points, and provide active guidance.

[0101] (2) Emotion combination 2: high negative emotion, low positive emotion

[0102] Condition: m({negative})≥0.7 and m({positive})<0.3.

[0103] Response strategy: Take emergency measures, issue official statements to clarify facts, suppress and guide negative public opinion, and strengthen communication with the public.

[0104] (3) Emotion combination 3: moderate positive and negative emotions

[0105] Conditions: 0.4≤m({positive})<0.7 and 0.4≤m({negative})<0.7.

[0106] Response strategy: Pay attention to public opinion trends, carry out appropriate positive publicity, respond to negative public opinion points, and maintain a balance in public opinion.

[0107] (4) Emotional combination 4: high-level emotion

[0108] Condition: m({neutral})≥0.5.

[0109] Response strategy: Strengthen policy publicity, guide public opinion in a positive direction, and ensure policy transparency and information disclosure.

[0110] Table 1 Evaluation response table

[0111]

[0112] In summary, the comprehensive public opinion identification and response decision-making method is an innovation of the public opinion management method based on evidence reasoning in this application. The public opinion management method based on evidence reasoning in this application is based on the comprehensive sentiment score, which can comprehensively evaluate the public opinion status and adopt corresponding response strategies according to different sentiment combinations. Through the analysis of the comprehensive sentiment score, the public opinion status can be accurately identified and the corresponding response strategy can be formulated to ensure the timeliness and effectiveness of public opinion management.

[0113] Compared with the existing public opinion analysis system, the public opinion management method based on evidence reasoning in this application has the following advantages:

[0114] 1. Comprehensive data collection: The public opinion management method based on evidence reasoning in this application can efficiently and legally collect data from multiple data sources (such as social media, news websites, forums, etc.) to ensure the comprehensiveness and diversity of public opinion information.

[0115] 2. Efficient data processing: Through automated data cleaning, deduplication and formatting, the public opinion management method based on evidence reasoning in this application can quickly process large-scale data and significantly improve data processing efficiency.

[0116] 3. Accurate sentiment analysis and topic identification: Using sentiment analysis models and topic generation models, the public opinion management method based on evidence reasoning in this application can accurately capture the sentiment tendencies and topic distribution in public opinion data, and provide more reliable analysis results.

[0117] 4. Effective evidence synthesis and conflict resolution: Through the Dempster-Shafer evidence theory, the public opinion management method based on evidence reasoning in this application can effectively synthesize evidence from different sources, resolve information conflicts, and provide reliable public opinion analysis results.

[0118] The public opinion management method based on evidence reasoning in this application involves a public opinion identification and response process based on evidence reasoning, specifically including: a method for collecting original public opinion data from multiple data sources; a method for cleaning, deduplicating, and formatting the collected original public opinion data to improve data quality; a method for using sentiment analysis models and topic generation models to score sentiment tendencies and classify topics for pre-processed public opinion data, and to correct sentiment scores accordingly; a method for synthesizing and resolving conflicts between evidence from different sources using Dempster-Shafer evidence theory; and a method for evaluating the public opinion status based on the synthesized sentiment scores and formulating corresponding response strategies. These methods all rely on computers and other devices with data processing or control functions to be implemented, and are an innovative process for processing public opinion data and controlling public opinion response strategies.

[0119] In an exemplary embodiment, Figure 2 As shown, the present application also provides a public opinion management device based on evidence reasoning, including:

[0120] 1) Data Collection Module:

[0121] Connection relationship: The data acquisition module is connected to the data preprocessing module.

[0122] Function: Collect raw public opinion data from multiple data sources (such as social media, news websites, forums, etc.), and use APIs and web crawlers to crawl data. Perform automatic crawling tasks at fixed preset intervals, and monitor public opinion events that require immediate response in real time.

[0123] 2) Data Preprocessing Module:

[0124] Connection relationship: The data preprocessing module is connected to the evidence collection and representation module.

[0125] Function: Clean, remove duplicates, and format the collected data to improve data quality. This includes data normalization, such as word segmentation, removal of stop words, and data standardization.

[0126] 3) Evidence Collection and Representation Module:

[0127] Connection relationship: The evidence collection and representation module is connected to the evidence synthesis and conflict resolution module.

[0128] Function: Use the sentiment analysis model to score the sentiment tendency of the preprocessed public opinion data and obtain the probability distribution of positive, negative and neutral sentiment. Use the subject generation model to classify the preprocessed public opinion data into topics and adjust the sentiment score according to the characteristics of the topic.

[0129] 4) Evidence Synthesis and Conflict Resolution Module:

[0130] Connection relationship: The evidence synthesis and conflict resolution module is connected to the public opinion status analysis and response strategy formulation module.

[0131] Function: Using the Dempster-Shafer evidence theory and the Dempster synthesis rule, it synthesizes evidence from multiple sources into a comprehensive piece of evidence and calculates the comprehensive sentiment score to resolve information conflicts and inconsistencies.

[0132] 5) Public Opinion Status Analysis and Response Strategy Formulation Module (Public Opinion Identification-Response Decision Module):

[0133] Connection relationship: The public opinion status analysis and response strategy formulation module is the final output module and is no longer connected to other modules.

[0134] Function: Based on the comprehensive sentiment score, evaluate the public opinion status and identify the overall sentiment tendency of public opinion, including the proportion of positive, negative and neutral sentiment. Different response strategies are adopted according to different combinations of comprehensive sentiment scores. Including formulating and implementing specific public opinion response plans to deal with different public opinion states.

[0135] In some embodiments, the present application also provides a computer program product, including a computer program, which implements the public opinion management method based on evidence reasoning when executed by a processor.

[0136] In some embodiments, the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the public opinion management method based on evidence reasoning.

[0137] In some embodiments, the present application also provides a computer device, including a processor, a memory, an input / output interface (I / O for short), a communication interface, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the public opinion management method based on evidence reasoning.

[0138] Among them, the processor, memory and 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 database of the computer device is used to store pending transactions. 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, the public opinion management method based on evidence reasoning is implemented.

[0139] It should be noted that the object information (including but not limited to object device information, object 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 object or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards.

[0140] Those skilled 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 (ReadOnlyMemory, 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. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. 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 are not limited to this.

[0141] The technical features of the above embodiments may be arbitrarily combined. 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.

[0142] 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 public opinion management method based on evidence reasoning, characterized in that: include: Obtain raw public opinion data from multiple data sources; Preprocessing the original public opinion data to obtain preprocessed public opinion data; Use sentiment analysis model and topic generation model to score the sentiment tendency of preprocessed public opinion data and obtain sentiment scores; Based on the sentiment score, the Dempster-Shafer evidence theory is used to obtain the comprehensive sentiment score; Analyze public opinion status and formulate response strategies based on comprehensive sentiment scores to complete public opinion management.

2. The public opinion management method based on evidence reasoning according to claim 1 is characterized in that: The obtaining of original public opinion data from multiple data sources specifically includes: Use the official API or data crawling tools of each data source to continuously obtain original public opinion data from multiple data sources at preset intervals, and record the timestamp and data ID of the captured original public opinion data.

3. The public opinion management method based on evidence reasoning according to claim 1 is characterized in that: The preprocessing of the original public opinion data to obtain the preprocessed public opinion data specifically includes: The original public opinion data is cleaned, deduplicated, standardized, segmented, noise filtered, stop word removed, and stemmed or morphologically restored to obtain the preprocessed public opinion data.

4. The public opinion management method based on evidence reasoning according to claim 1 is characterized in that: The sentiment analysis model and the topic generation model are used to score the sentiment tendency of the preprocessed public opinion data to obtain the sentiment score, which specifically includes: Use the sentiment analysis model to score the sentiment tendency of the preprocessed public opinion data and obtain the initial sentiment score Among them, s pos 、s neg and neu They represent the initial positive, negative, and neutral sentiment scores of the preprocessed public opinion data respectively; Use the topic generation model to classify the preprocessed public opinion data into topics and obtain topic scores Among them, l i represents the probability of the ith topic of the preprocessed public opinion data, and k is the total number of topics; The initial sentiment score is modified based on the topic score to obtain the sentiment score.

5. The public opinion management method based on evidence reasoning according to claim 4 is characterized in that: The initial sentiment score is corrected based on the topic score to obtain the sentiment score, specifically including: Based on topic score Use the formula and For pos 、s neg and neu Correction is performed to obtain the corrected positive, negative and neutral sentiment scores s' pos 、s' neg and s' neu ; Among them, α i is the positive sentiment weight of the ith topic; β i is the negative sentiment weight of the ith topic; γ i is the neutral sentiment weight of the i-th topic; According to the corrected positive, negative and neutral sentiment scores s' pos 、s' neg and s' neu Get sentiment score 6. The public opinion management method based on evidence reasoning according to claim 5 is characterized in that: Based on the sentiment score, the Dempster-Shafer evidence theory is used to obtain a comprehensive sentiment score, which specifically includes: Based on the sentiment score of the nth preprocessed public opinion data Using formula m n ({positive})=s' pos,n 、m n ({negative})=s' neg,n and m n ({neutral})=s' neu,n The corrected positive, negative and neutral sentiment scores s' for the nth preprocessed public opinion data pos,n 、s' neg,n and s' neu,n Mapping is performed to obtain the sentiment score m of the nth pre-processed public opinion data after mapping n =(m n ({positive}),m n ({negative}),m n ({neutral})); where m n ({positive}), m n ({negative}) and m n ({neutral}) are the positive, negative, and neutral sentiment scores of the nth preprocessed public opinion data after mapping; Based on m n =(m n ({positive}),m n ({negative}),m n ({neutral})), and respectively adopt the formulas through Dempster-Shafer evidence theory m2({positive}))、 and The comprehensive positive emotion score m({positive}), the comprehensive negative emotion score m({negative}) and the comprehensive neutral emotion score m({neutral}) are calculated; where K is the conflict coefficient; n=1, 2; Based on m({positive}), m({negative}) and m({neutral}), we get the comprehensive sentiment score m=(m({positive}),m({negative}),m({neutral})).

7. The method for public opinion management based on evidence reasoning according to claim 6 is characterized in that: The above mentioned public opinion status analysis and response strategy formulation based on the comprehensive sentiment score completes public opinion management, specifically including: When the comprehensive positive sentiment score m({positive}) is greater than or equal to 0.7, the public opinion status is generally positive; when the comprehensive positive sentiment score m({positive}) is greater than or equal to 0.4 and less than 0.7, the public opinion status is relatively positive; when the comprehensive positive sentiment score m({positive}) is less than 0.4, the public opinion status is general; When the comprehensive negative sentiment score m({negative}) is greater than or equal to 0.7, the public opinion status is generally negative; when the comprehensive negative sentiment score m({positive}) is greater than or equal to 0.4 and less than 0.7, the public opinion status is relatively negative; when the comprehensive negative sentiment score m({positive}) is less than 0.4, the public opinion status is general; When the comprehensive neutral sentiment score m({neutral}) is greater than or equal to 0.5, the public opinion state is neutral; when the comprehensive neutral sentiment score m({neutral}) is less than 0.5, the public opinion state is non-neutral; Formulate response strategies based on different combinations of public opinion status to complete public opinion management.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for public opinion management based on evidence reasoning described in any one of claims 1 to 7 is implemented.

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 method for public opinion management based on evidence reasoning described in any one of claims 1 to 7 is implemented.

10. 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 public opinion management method based on evidence reasoning as described in any one of claims 1 to 7.

Images