An artificial intelligence-based network public opinion risk monitoring and early warning system and method

Abstract

The present invention discloses an artificial intelligence-based network public opinion risk monitoring and early warning system and method. The system includes: an application subsystem, a client terminal, and a storage subsystem, wherein: the application subsystem includes: a network public opinion risk ontology construction module, an index system construction and prediction module, an intelligent monitoring module, a scenario modeling and simulation module, a risk level determination module, an impact assessment generation module, and an automatic early warning module. Using the present invention facilitates the realization of real-time intelligent monitoring and early warning of network public opinion risks, improves the timeliness and accuracy of network public opinion risk monitoring and early warning, so as to cope with challenges such as large amounts of monitoring data, strong timeliness, and high professional requirements, and provides strong support for network public opinion risk governance.

Description

Technical Field

[0001] The present invention relates to the technical field of network public opinion risk monitoring, and more specifically, to a network public opinion risk monitoring and early warning system and method based on artificial intelligence. Background Art

[0002] Network public opinion emerges due to the occurrence of an event, takes the Internet platform as a carrier, and includes the collective understanding, attitude, and emotional behavior tendency of Internet users towards this event. With the development of technologies such as the Internet and artificial intelligence, network public opinion has characteristics such as richer content, faster dissemination, more confusing topics, and mutual reflection with real life.

[0003] Although there are currently many studies and solutions for network public opinion analysis and early warning, it still mainly relies on the manual tracking of public opinion analysts to judge the status and influence of network public opinion; however, in the era of intelligence, network public opinion has problems such as a large amount of information, strong timeliness, and high professional requirements, and aspects such as the timeliness and accuracy of the risk monitoring and early warning plan need to be further improved; in order to avoid the impact caused by network public opinion, more attention needs to be paid to the real-time monitoring, in-depth evaluation, timely early warning, and rapid response of network public opinion risks. Therefore, real-time intelligent and accurate monitoring and early warning of network public opinion risks have become a problem that must be solved at present. Summary of the Invention

[0004] In view of this, the present invention provides a network public opinion risk monitoring and early warning system and method based on artificial intelligence that at least solves the above-mentioned partial technical problems, which is convenient for realizing real-time intelligent monitoring and early warning of network public opinion risks, improves the timeliness and accuracy of network public opinion risk monitoring and early warning, and responds to challenges such as a large amount of monitoring data, strong timeliness, and high professional requirements, providing strong support for network public opinion risk governance.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] In the first aspect, the present invention provides a network public opinion risk monitoring and early warning system based on artificial intelligence, and the system includes: an application subsystem, a client terminal, and a storage subsystem, wherein:

[0007] The application subsystem includes: a network public opinion risk ontology construction module, an index system construction and prediction module, an intelligent monitoring module, a scenario modeling and simulation module, a risk level determination module, an impact assessment generation module, and an automatic early warning module; wherein:

[0008] The network public opinion risk ontology construction module is used to construct a network public opinion risk ontology model by using an ontology modeling assistance method, and formally express the concepts and relationships of the subject, type, propagation law, influencing factors, and coping strategies of network public opinion.

[0009] The index system construction and prediction module is used to construct an online public opinion risk assessment index system based on the online public opinion risk ontology and in combination with data related to online public opinion evaluation obtained from the Internet. At the same time, preprocess the obtained data, and calculate the measure values of the online public opinion risk assessment indexes of each data in combination with the entropy weight method; predict the future trend of the indexes based on the measured data;

[0010] The intelligent monitoring module is used to monitor online public opinion risk information. Based on the monitoring information, automatically capture hot risk events in real time, and determine the fields, associated indexes, and associated events to which the hot risk events belong;

[0011] The scenario modeling and simulation module is used to classify online public opinion risk scenarios, clarify typical risk scenarios, and construct an online public opinion risk scenario model through a system dynamics model; it is also used to simulate the risk scenario model, analyze the impact of risk events on associated indexes and the impact of response strategies on online public opinion;

[0012] The risk level determination module is used to determine the risk levels of associated indexes under the influence of hot risk events based on the simulation of the risk scenario model, and determine the event risk level through the risk levels of the associated indexes;

[0013] The impact assessment generation module is used to generate an assessment text of the impact of online public opinion hot risk events based on the descriptions of hot risk events, the impact on associated indexes, and information on event risk levels, using a large language model;

[0014] The automatic warning module is used to automatically analyze the fields and event contents of hot risk events, and generate a warning report for hot risk events in combination with the assessment text of the impact of hot risk events, and send the warning report to the customer terminal in real time;

[0015] The customer terminal is used to receive the warning report sent by the automatic warning module and push warning information to the customer;

[0016] The storage subsystem includes: an ontology library for storing the online public opinion risk ontology model, an index library for storing the online public opinion risk assessment index system, a scenario model library for storing the online public opinion risk scenario model, an event library for storing typical hot online public opinion risk events, a database for storing online public opinion risk information data, and a large language model storage library.

[0017] Further, in the online public opinion risk ontology construction module, the ontology modeling auxiliary methods for constructing the online public opinion risk ontology model include: data capture, term extraction, relationship extraction, and text analysis.

[0018] Furthermore, in the index system construction and prediction module, the preprocessing of the obtained data includes: data cleaning, index positive transformation, and normalization.

[0019] Furthermore, in the index system construction and prediction module, the measure values of the network public opinion risk assessment indexes for each data are calculated by combining the entropy weight method. The calculation formula includes:

[0020]

[0021] where i represents the i-th index, with a total of n indexes, j represents the j-th evaluation sample of the index, with a total of m evaluation samples, and e i represents the entropy value of the i-th index X i , p ij represents the proportion of the j-th evaluation sample x ij of the i-th index after normalization, ω i represents the weight value of the i-th index, k j represents the target measure value after the fusion of the j-th evaluation sample index.

[0022] Furthermore, in the intelligent monitoring module, the methods of data crawling and multi-modal data analysis are adopted to automatically discover and capture network public opinion hot spot risk events, and the monitoring information includes text and pictures.

[0023] Furthermore, in the scenario modeling and simulation module, the system dynamics model is:

[0024]

[0025] where x i (k) represents the risk level of the i-th index at the k-th stage, represents the comprehensive effect of the risk levels of all indexes directly associated with the i-th index on it, N i(k) represents the set of all indexes directly affecting the i-th index, α ji represents the transfer coefficient of the influence effect of the associated index j on the i-th index, R ji (k) represents the influence degree of the i-th index on the i-th index, ΔR ji (k) represents the interaction intensity between indexes, ΔE i (k) represents the comprehensive intervention effect of the intervention strategy on the index, β i is the influence intensity of the event on the index, S i(k) represents the set of all strategies directly associated with the i-th index, =γ li x i (k) represents the effect of suppressing the index risk, γ li is the influence intensity of the strategy on the index, noise i (k) represents the noise value.

[0026] Furthermore, the application subsystem further includes an expert judgment module, which is used to provide a window for experts to receive and view hot risk events in real time, select typical hot risk events, enter the network public opinion risk scenario modeling and simulation module, improve the network public opinion risk scenario model, modify and improve the risk level and impact assessment of typical hot events, and form a warning report with the coping strategies proposed by experts for typical hot risk events.

[0027] In a second aspect, the present invention also provides an artificial intelligence-based network public opinion risk monitoring and warning method, which is applied to the above-mentioned artificial intelligence-based network public opinion risk monitoring and warning system for network public opinion risk monitoring and warning. The method includes:

[0028] Using the ontology modeling assistance method, construct a network public opinion risk ontology model to formally express the concepts and relationships of the subjects, types, propagation laws, influencing factors, and coping strategies of network public opinion.

[0029] Based on the network public opinion risk ontology, combined with the network public opinion evaluation-related data obtained from the Internet, construct a network public opinion risk assessment index system; at the same time, preprocess the obtained data, and calculate the measurement values of the network public opinion risk assessment indexes of each data in combination with the entropy weight method; use the ARIMA method to automatically predict the future trend of the indexes based on the historical measurement data.

[0030] Monitor network public opinion risk information, and based on the monitoring information, automatically capture hot risk events in real time, and determine the fields, associated indicators, and associated events to which the hot risk events belong.

[0031] Conduct network public opinion risk scenario classification, clarify typical risk scenarios, and construct a network public opinion risk scenario model through a system dynamics model; it is also used for simulating the risk scenario model to analyze the impact of risk events on associated indicators and the impact of coping strategies on network public opinion.

[0032] Based on the simulation of the risk scenario model, determine the risk level of the associated indicators under the influence of hot risk events, and determine the event risk level through the risk level of the associated indicators.

[0033] Based on the information of the hot risk event description, the impact on the associated indicators, and the event risk level, use a large language model to generate an evaluation text on the impact of the network public opinion hot risk event.

[0034] Automatically analyze the field and content of the hot risk event, and generate a hot risk event warning report in combination with the evaluation text on the impact of the hot risk event, and send the warning report to the customer terminal in real time.

[0035] The customer terminal receives the sent warning report and pushes the warning information to the customer;

[0036] Store the network public opinion risk ontology model, network public opinion risk assessment index system, network public opinion risk scenario model, typical hot network public opinion risk events, network public opinion risk information data, and large language model.

[0037] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0038] The present invention provides a network public opinion risk monitoring and warning system and method based on artificial intelligence, which facilitates the realization of real-time intelligent monitoring and warning of network public opinion risks, improves the timeliness and accuracy of network public opinion risk monitoring and warning, and copes with challenges such as large amounts of monitoring data, strong timeliness, and high professional requirements, providing strong support for network public opinion risk governance.

[0039] Other features and advantages of the present invention will be described in the following specification, and part of them will become obvious from the specification or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in the written specification and the accompanying drawings.

[0040] The technical solutions of the present invention will be further described in detail below through the accompanying drawings and embodiments. Description of the Drawings

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other accompanying drawings can be obtained based on these drawings without creative efforts.

[0042] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation to the present invention.

[0043] Figure 1 It is a schematic structural diagram of a network public opinion risk monitoring and warning system based on artificial intelligence provided by an embodiment of the present invention.

[0044] Figure 2 It is a schematic flowchart of a network public opinion risk monitoring and warning method based on artificial intelligence provided by an embodiment of the present invention. Detailed Embodiments

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. The technical solutions of the present invention will be described clearly and completely. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0046] See Figure 1 As shown, the embodiments of the present invention provide an artificial intelligence-based network public opinion risk monitoring and early warning system, which includes three parts: an application subsystem, a client terminal, and a storage subsystem; wherein:

[0047] In the embodiments of the present invention, the application subsystem includes: a network public opinion risk ontology construction module, an index system construction and prediction module, an intelligent monitoring module, a scenario modeling and simulation module, a risk level determination module, an impact assessment generation module, and an automatic early warning module;

[0048] The storage subsystem includes: an ontology library for storing network public opinion risk ontology models, an index library for storing network public opinion risk assessment index systems, a scenario model library for storing network public opinion risk scenario models, an event library for storing typical hot spot risk events of network public opinion, a database for storing network public opinion risk information data, and a large language model storage library.

[0049] The module functions of the application subsystem of the present invention are introduced below:

[0050] 1. Ontology Construction Module

[0051] It has ontology modeling assistance functions such as network public opinion information capture, term extraction, and relationship extraction, can carry out visual modeling of network public opinion risk ontology, and supports model analysis and verification.

[0052] 2. Index System Construction and Prediction Module

[0053] It is used to visually display the structure of the index system and the relationships between indexes, measure data, combine methods such as index information entropy quantization and expert evaluation, automatically predict the future trends of indexes, support dynamic update of measure data, and can store information such as evaluation indexes, measures, data, and sources.

[0054] 3. Intelligent Monitoring Module

[0055] It is used for customizing network public opinion risk monitoring information, and based on the monitoring information, it can automatically capture hot spot risk events in real time and automatically determine the fields, associated indexes, and associated events to which the hot spot risk events belong.

[0056] 4. Scenario Modeling and Simulation Module

[0057] Based on modeling methods such as system dynamics, construct a visual network public opinion risk scenario model. Through the simulation of the risk scenario model, analyze the impact of hot risk events on related indicators and the impact of response strategies on network public opinion, and store the network public opinion risk model to form a network public opinion risk scenario model library.

[0058] 5. Risk Level Judgment Module

[0059] Intelligently match the network public opinion scenario model of hot risk events, vectorize the titles and contents of hot events, calculate the distance from the feature vectors in the scenario model feature library, obtain similarity evaluation indicators, and use the scenario model with the highest similarity index as the simulation calculation model for the event. Subsequently, based on the risk scenario model simulation, analyze the impact of hot risk events on their core related indicators, compare the impact of events on indicators with the results of automated indicator trend prediction, determine the risk level of related indicators under the influence of hot risk events, and combine the risk levels of related indicators to determine the event risk level.

[0060] 6. Impact Assessment Intelligent Generation Module

[0061] Based on information such as the description of hot risk events, the impact on related indicators, and the event risk level, use a large language model to generate a text for assessing the impact of network public opinion hot risk events.

[0062] 7. Automatic Early Warning Module

[0063] Automatically analyze the field and content of hot risk events to determine the recipients of early warnings, and send the early warning reports of hot risk events to customer terminals in real time.

[0064] In a preferred embodiment, the application subsystem of the system further includes: an expert judgment module, which is used to provide an expert with a window for receiving and viewing hot risk events in real time, select typical hot risk events, enter the network public opinion risk scenario modeling and simulation module, improve the network public opinion risk scenario model, modify and improve the risk level and impact assessment of typical hot events, and is used to form an early warning report based on the response strategies proposed by experts for typical hot risk events.

[0065] The following details the specific implementation manners of the present invention:

[0066] (1) Construction of Network Public Opinion Risk Ontology.

[0067] In this embodiment, using keywords such as network public opinion risk, public opinion risk, and network opinion risk, based on technologies such as data capture, term extraction, relationship extraction, and text analysis, relevant literature, reports, etc. on network public opinion analysis methods and practices on Web of Science and CNKI are obtained and analyzed to identify the core concepts and their relationships such as the types, subjects, dissemination laws, risk management, influencing factors, and coping strategies of network public opinion. In the ontology construction module, a network public opinion risk ontology model is constructed.

[0068] (2) Construction of the network public opinion risk assessment index system.

[0069] 2.1: Based on the topic analysis of network public opinion risk, it is found that network public opinion risk involves different fields such as economy and technology, so the first-level indicators of economy and technology can be defined.

[0070] 2.2: Through the detailed analysis of network public opinion risk under the first-level indicators, the secondary evaluation indicators are clarified. For example, the secondary indicators under the first-level economic evaluation indicator include: industrial scale benefit, capital market activity, industrial development vitality, enterprise competitiveness, R & D expenditure, etc.

[0071] 2.3: Based on the collected evaluation data, the measurement names and measurement values of the secondary evaluation indicators are clarified. For example, enterprise competitiveness includes: enterprise market value, product market share, number of products, number of employees, etc. In addition, for some measurements, the measurement values need to be calculated based on data. At this time, data preprocessing work such as data cleaning, index positive transformation, and normalization is first performed on the underlying data, and then the weight of the influence of each data on the index measurement is estimated by methods such as the entropy weight method. After correction, the weight is determined, integrated into the measurement value, and the data source, update cycle, etc. are clarified.

[0072] Among them, data normalization refers to removing the large numerical differences caused by the influence of different data dimensions or dimension units, and mapping the data to a specified range. The present invention uses Min - Max normalization (also known as deviation standardization, as shown in the following formula) to perform a linear transformation on the original data and map the result to between 0 and 1.

[0073]

[0074] Among them, X' represents the dimensionless data, X represents the observed data, X min represents the minimum value of the data, and X max represents the maximum value of the data.

[0075] Data positive transformation is to judge whether the larger the data value is, the better. If so, it is positive data; if not, it is reverse data. For positive data, the above formula is used for measurement normalization, and the normalization of reverse data is shown in the following formula.

[0076]

[0077] For the index measurement calculated from multiple data, the present invention mainly uses the entropy weight method to determine the weights of each data, mainly with the objective weighting method and corrected by the subjective weighting method. The weights are determined according to the information entropy of each data. According to information theory, entropy is a measure of the disorder degree of a system. The entropy value e i of the i-th data X i is calculated as shown in the following formula.

[0078]

[0079] where i represents the i-th index, with a total of n indexes, j represents the j-th evaluation sample of the index, with a total of m evaluation samples, and e i represents the entropy value of the i-th index X i , p ij represents the proportion of the j-th evaluation sample x ij of the i-th index after normalization.

[0080]

[0081] Furthermore, the entropy of all data is normalized. If the entropy value of a certain data is large, it means that the data has a large degree of dispersion, which can be regarded as the data containing richer information. Therefore, when comprehensively evaluating the measurement, the influence (i.e., weight) of this data is greater. The objective calculation formula of each index weight ω i is as follows:

[0082]

[0083] Furthermore, after preprocessing such as positive transformation and normalization of the underlying data, and calculating the data according to the following formula, the target measure k j value after fusing the indexes of the j-th evaluation sample can be obtained:

[0084]

[0085] 2.4: Analyze the relationship between the secondary evaluation indexes, and clarify their positive or negative impacts. For example, the higher the activity of the capital market, the higher the vitality of industrial development. In the index system construction and prediction module, construct the index system.

[0086] 2.5: Store the evaluation indexes, measures, and data.

[0087] (3) Prediction of the trend of the network public opinion risk assessment index.

[0088] In the index system construction and prediction module, based on the measured values of the network public opinion risk assessment indicators, methods such as ARIMA (Auto Regressive Integrated Moving Average) can be used to automatically predict the future trends of the indicators based on the historical measured data and give the estimated value range of the confidence interval.

[0089] (4) Intelligent monitoring of network public opinion.

[0090] 4.1: In the intelligent monitoring module, based on the network public opinion risk ontology and risk assessment index system, determine information such as the fields, subjects, keywords, languages, and main sources of network public opinion risk monitoring. Among them, the monitoring fields include economic, scientific and technological, etc. fields. Separate subjects of concern are set for each field.

[0091] 4.2: Adopt technologies such as data crawling and multi-modal data analysis to automatically discover and capture network public opinion hot spot risk events, including texts and pictures, etc., and based on large language model technology, realize language translation and intelligently generate brief introductions of risk events; the above large language model technology, etc. are existing mature technologies and will not be elaborated here.

[0092] 4.3: Based on technologies such as neural network models, realize event feature extraction and classification, and determine the fields to which the hot spot risk events belong.

[0093] 4.4: Retrieve the network public opinion risk assessment index system and determine the core correlation indicators of the hot spot risk events based on text similarity calculation.

[0094] 4.5: Automatically capture the associated events of the hot spot risk events based on text similarity calculation.

[0095] (5) Construction of the network public opinion risk scenario model library.

[0096] 5.1: Based on the analysis of network public opinion risk monitoring events, conduct network public opinion risk scenario classification and clarify typical risk scenarios.

[0097] 5.2: In the network public opinion risk scenario modeling and simulation module, based on the network public opinion risk ontology, risk assessment index system, etc., combined with modeling methods such as system dynamics, construct a visual network public opinion risk scenario model.

[0098] For example, regard the network public opinion risk as a complex system, the risk assessment index system as the system state, the risk events as external disturbances, the intervention strategies as system inputs, and finally transform the original problem into a control problem to construct a system dynamics model for general risk assessment of network public opinion; among them, the total system of equations is as follows,

[0099]

[0100] Among them, x i (k) represents the risk level of the k-th stage of index i, and its value is obtained by comprehensively calculating its underlying data sources. represents the comprehensive effect of the risk levels of all indexes directly associated with index i. N i(k) represents the set of all indexes that directly affect index i. α ji represents the transfer coefficient of the influence effect of associated index j on index i. The higher this coefficient, the higher the degree of association between the two indexes. R ji (k) represents the influence degree of index i on index i, and its value is positively correlated with the risk level of index i itself. That is, the higher the risk of index i, the higher the additional risk it transfers to index i. ΔR ji (k) = x i (k)(x j (k) - x j _base) / x j _base represents the interaction strength between indexes. ΔE i (k) = e -θk β i x i (k) represents the comprehensive intervention effect of the intervention strategy on the index. β i is the influence strength of the event on the index. S i(k) represents the set of all strategies directly associated with index i. ΔS li (k) = γ li x i (k) has the effect of suppressing the index risk. γ li is the influence strength of the strategy on the index. noise i (k) represents the noise value.

[0101] Furthermore, various network public opinion risk scenario models form a network public opinion risk scenario model library.

[0102] (6) Intelligent determination of the risk level of network public opinion hot events. In the intelligent determination module of the hot event risk level:

[0103] 6.1: Intelligently match the network public opinion scenario model of the hot event risk, vectorize the title and content of the hot event, calculate the distance from the feature vectors in the scenario model feature library, obtain the similarity evaluation index, and use the scenario model with the highest similarity index as the simulation calculation model of this event.

[0104] 6.2: Through the simulation of the risk scenario model, analyze the impact of the hot event risk on its core associated indexes.

[0105] 6.3: Compare the impacts of hot-spot risk events on indicators and the predicted results of indicator trends to determine the risk levels of associated indicators under the influence of hot-spot risk events.

[0106] 6.4: Combine the risk levels of associated indicators with the risk levels of similar events in the event library to determine the event risk level.

[0107] (7) Intelligent generation of the impact assessment of online public opinion hot-spot events. In the intelligent generation module of the hot-spot risk event impact assessment:

[0108] Based on information such as the description of hot-spot risk events, their impacts on associated indicators, risk levels, and the impact assessments of similar events in the event library, use large language models to generate the impact assessment of online public opinion hot-spot risk events.

[0109] (8) Human-machine collaborative judgment of typical hot-spot events in online public opinion.

[0110] 8.1: In the expert judgment module, it is used to provide experts with a window to receive and view hot-spot risk events in real time, displaying the event list; experts can receive and view the intelligent risk level evaluation and impact assessment results of hot-spot risk events in real time, which is convenient for selecting typical hot-spot risk events. Information such as event ID, title, field, event summary, event time, updated event, original link, status, online time, edit link, etc. is displayed in the event list.

[0111] 8.2: Based on the analysis of typical hot-spot risk events and their associated events, modify and improve the online public opinion risk scenario model.

[0112] 8.3: In the details of hot-spot risk events, confirm the risk level of typical hot-spot risk events and improve the text of the impact assessment of hot-spot risk events.

[0113] 8.4: In the details of hot-spot risk events, put forward corresponding countermeasure suggestions for the impacts of typical hot-spot risk events;

[0114] 8.5: In the details of hot-spot risk events, form a warning report for typical hot-spot risk events.

[0115] (9) Automatic warning of typical hot-spot events in online public opinion.

[0116] 9.1: Determine the warning recipients according to the automated analysis of the fields and event contents of typical hot-spot risk events.

[0117] 9.2: With the help of the early warning function and the client terminal, the early warning information of typical hot - spot risk events is sent to the client in real - time, supporting the client to grasp the public opinion situation in real - time. In the automatic early warning module, record the number of the latest dynamic events every day, the distribution of candidate event fields every day, and the change trend of the number of typical hot - spot events that have been pushed. And display information such as the ID, title, field, event summary, event time, updated event, original text link, status, online time, editing link, push terminal, etc. of the candidate events and the pushed events.

[0118] 9.3: Customers can provide suggestions on the public opinion risk monitoring and early warning work in real - time through the early warning client terminal, so as to continuously improve the quality of network public opinion risk monitoring and early warning work.

[0119] (10) Construction of the network public opinion event library.

[0120] Based on the typical hot - spot risk events in the network public opinion risk monitoring and early warning, construct a network public opinion risk event library, record the hot - spot risk events of network public opinion, including event time, summary, subject, field, related indicators, related events, risk level, impact assessment, response strategies, etc., providing reference for the intelligent monitoring and early warning of continuous network public opinion hot - spot risk events.

[0121] From the description of the above embodiments, those skilled in the art can know that the present invention provides a network public opinion risk monitoring and early warning system based on artificial intelligence. The system includes: an application subsystem, a client terminal, and a storage subsystem. The present invention integrates methods and technologies such as ontology construction, large - language models, generative artificial intelligence, system modeling, and expert judgment based on Internet information, constructs a network public opinion risk ontology and a risk assessment index system, supports intelligent monitoring of network public opinion risks, risk scenario modeling, professional risk judgment and impact assessment, and real - time early warning, improving the timeliness and accuracy of network public opinion risk monitoring and early warning to cope with challenges such as large amounts of monitoring data, strong timeliness, and high professional requirements, providing strong support for network public opinion risk governance. The specific advantages are as follows:

[0122] ①. To address challenges such as large amounts of data in the network public opinion field and low efficiency of manually constructing ontologies, based on ontology modeling assistance technology, construct a network public opinion risk ontology model to formally express the core concepts and their relationships in this field, which is not only conducive to the communication and understanding of knowledge in the network public opinion risk field, but also provides important support for the construction of the risk assessment index system, risk scenario modeling, large - language model construction, and risk event monitoring in this field.

[0123] ②. Through the network public opinion risk assessment index system, classify and hierarchically evaluate network public opinion risks, comprehensively and quickly measure the network public opinion risk situation, assist in accurately positioning and analyzing problems, and support network public opinion risk assessment.

[0124] ③. Through customized real-time intelligent monitoring of online public opinion risks, quickly and accurately capture hot risk events of online public opinion, reduce labor costs, and improve information monitoring efficiency.

[0125] ④. Through the modeling and simulation of online public opinion risk scenarios, realize the visual modeling, mechanism analysis and evolutionary reasoning of complex online public opinion risk scenarios, and support in-depth analysis of typical online public opinion risk scenarios and risk events.

[0126] ⑤. The intelligent determination of the risk level of hot events and the intelligent generation of impact assessment based on the event library, large language model, etc. will effectively improve the efficiency of risk determination and impact assessment of hot events, and provide important references for expert judgment.

[0127] ⑥. Through automatic early warning and customer terminals, realize the real-time sending of early warning information to support customers' rapid response. At the same time, receive customer feedback information in real time to promote the continuous update of the risk monitoring and early warning system.

[0128] Furthermore, the embodiment of the present invention also provides an online public opinion risk monitoring and early warning method based on artificial intelligence, which is applied to an online public opinion risk monitoring and early warning system based on artificial intelligence in the above embodiment for human-machine collaborative online public opinion risk monitoring and early warning based on artificial intelligence. The specific process is shown in Figure 2 as follows. The method includes:

[0129] S1. Use the ontology modeling assistance method to construct an online public opinion risk ontology model, and formally express the concepts and relationships of the subjects, types, propagation laws, influencing factors and coping strategies of online public opinion.

[0130] S2. Based on the online public opinion risk ontology, combined with the data related to online public opinion evaluation obtained from the Internet, construct an online public opinion risk assessment index system; at the same time, preprocess the obtained data, and calculate the measure values of the online public opinion risk assessment indicators of each data in combination with the entropy weight method; use the ARIMA method to automatically predict the future trend of the indicators based on the historical measure data.

[0131] S3. Monitor online public opinion risk information, and based on the monitoring information, automatically capture hot risk events in real time, and determine the fields, associated indicators and associated events to which the hot risk events belong.

[0132] S4. Classify online public opinion risk scenarios, clarify typical risk scenarios, and construct an online public opinion risk scenario model through a system dynamics model; it is also used for the simulation of the risk scenario model to analyze the impact of risk events on associated indicators and the impact of coping strategies on online public opinion.

[0133] S5. Based on the risk scenario model simulation, determine the risk levels of associated indicators under the influence of hot-spot risk events, and determine the event risk level through the risk levels of the associated indicators;

[0134] S6. Based on the description of the hot-spot risk event, its impact on associated indicators, and the information on the event risk level, use a large language model to generate an evaluation text on the impact of the online public opinion hot-spot risk event;

[0135] S7. Automatically analyze the field and content of the hot-spot risk event, and combine it with the evaluation text on the impact of the hot-spot risk event to generate a warning report on the hot-spot risk event, and send the warning report to the customer terminal in real time;

[0136] S8. The customer terminal receives the sent warning report and pushes warning information to the customer;

[0137] S9. Store the online public opinion risk ontology model, the online public opinion risk assessment index system, the online public opinion risk scenario model, the typical online public opinion hot-spot risk events, the online public opinion risk information data, and the large language model.

[0138] For the method for monitoring and warning online public opinion risks based on artificial intelligence provided by the embodiments of the present invention, its implementation principle and the technical effects produced are the same as those of the foregoing system embodiments. For the sake of brief description, for the parts not mentioned in this embodiment, please also refer to the corresponding content in the foregoing system embodiments, which will not be elaborated here.

[0139] In addition, the embodiments of the present invention also provide a storage medium, on which one or more programs readable by a computing device are stored. The one or more programs include instructions that, when executed by the computing device, cause the computing device to execute the method for monitoring and warning online public opinion risks based on artificial intelligence in the foregoing embodiments.

[0140] In the embodiments of the present invention, the storage medium may be, for example, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the above. More specific examples of the storage medium (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, and any suitable combination of the above.

[0141] Those skilled in the art should understand that the embodiments of the present invention can be provided as a system, a method, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0142] It should be noted that the word "comprising" does not exclude the presence of components not listed in the claims. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The present invention can be implemented by means of hardware including several different components and by means of a suitably programmed computer.

[0143] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other.

[0144] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. An artificial intelligence-based network public opinion risk monitoring and early warning system, characterized in that, The system includes: an application subsystem, a customer terminal, and a storage subsystem, where: The application subsystem includes: an online public opinion risk ontology construction module, an index system construction and prediction module, an intelligent monitoring module, a scenario modeling and simulation module, a risk level determination module, an impact assessment generation module, and an automatic warning module; where: The online public opinion risk ontology construction module is used to construct an online public opinion risk ontology model by using an ontology modeling assistance method, and formally express the concepts and relationships of the main body, type, propagation law, influencing factors, and coping strategies of online public opinion. The index system construction and prediction module is used to construct an online public opinion risk assessment index system based on the online public opinion risk ontology and in combination with the online public opinion evaluation-related data obtained from the Internet; at the same time, preprocess the obtained data, and calculate and obtain the measure values of the online public opinion risk assessment indexes of each data in combination with the entropy weight method; predict the future trend of the indexes based on the measured data. The intelligent monitoring module is used to monitor online public opinion risk information, and based on the monitoring information, automatically capture hot risk events in real time, and determine the fields, associated indexes, and associated events to which the hot risk events belong. The scenario modeling and simulation module is used to classify online public opinion risk scenarios, clarify typical risk scenarios, and construct an online public opinion risk scenario model through a system dynamics model; it is also used to simulate the risk scenario model, and analyze the impact of risk events on associated indexes and the impact of coping strategies on online public opinion. The risk level determination module is used to determine the risk levels of associated indexes under the influence of hot risk events based on the simulation of the risk scenario model, and determine the event risk level through the risk levels of the associated indexes. The impact assessment generation module is used to generate an assessment text of the impact of online public opinion hot risk events by using a large language model based on the description of hot risk events, the impact on associated indexes, and the information of the event risk level. The automatic warning module is used to automatically analyze the fields and event contents to which hot risk events belong, and generate a warning report of hot risk events in combination with the assessment text of the impact of hot risk events, and send the warning report to the customer terminal in real time. The customer terminal is used to receive the warning report sent by the automatic warning module and push warning information to the customer. The storage subsystem includes: an ontology library for storing the online public opinion risk ontology model, an index library for storing the online public opinion risk assessment index system, a scenario model library for storing the online public opinion risk scenario model, an event library for storing typical hot risk events of online public opinion, a database for storing online public opinion risk information data, and a large language model storage library.

2. The network public opinion risk monitoring and early warning system based on artificial intelligence according to claim 1, characterized in that, In the online public opinion risk ontology construction module, the ontology modeling assistance method for constructing the online public opinion risk ontology model includes: data capture, term extraction, relationship extraction, and text analysis.

3. An artificial intelligence-based network public opinion risk monitoring and early warning system according to claim 1, characterized in that, In the index system construction and prediction module, the preprocessing of the obtained data includes: data cleaning, index positive transformation, and normalization.

4. An artificial intelligence-based network public opinion risk monitoring and early warning system according to claim 1, characterized in that, In the index system construction and prediction module, the measure values of the online public opinion risk assessment indexes of each data are calculated in combination with the entropy weight method, and the calculation formula includes: Among them, i represents the i-th index, with a total of n indices, and j represents the j-th evaluation sample of the index, with a total of m evaluation samples, e i represents the entropy value of the i-th index X i . p ij represents the proportion of the j-th evaluation sample of the i-th index after normalization, x ij . ω i represents the weight value of the i-th index, k j represents the target measure value after the fusion of the j-th evaluation sample index.

5. An artificial intelligence-based online public opinion risk monitoring and early warning system according to claim 1, characterized in that, In the intelligent monitoring module, the methods of data crawler and multi-modal data analysis are adopted to automatically discover and capture hot risk events of online public opinion, and the monitoring information includes text and pictures.

6. An artificial intelligence-based online public opinion risk monitoring and early warning system according to claim 1, characterized in that, In the scenario modeling and simulation module, the system dynamics model is as follows: Among them, x i (k) represents the risk level of the k-th stage of index i, represents the comprehensive effect of the risk levels of all indexes directly associated with index i, N i(k) represents the set of all indexes directly affecting index i, α ji represents the transfer coefficient of the influence effect of associated index j on index i, R ji (k) represents the influence degree of index i on index i, ΔR ji (k) represents the interaction strength between indexes, ΔE i (k) represents the comprehensive intervention effect of the intervention strategy on the index, β i is the influence strength of the event on the index, S i(k) represents the set of all strategies directly associated with index i, ΔS li (k) = γ li x i (k) represents the effect of suppressing the index risk, γ li is the influence strength of the strategy on the index, noise i (k) represents the noise value.

7. An artificial intelligence-based network public opinion risk monitoring and early warning system according to claim 1, characterized in that, The application subsystem further includes an expert judgment module, which is used to provide an expert with a window for receiving and viewing hot risk events in real time, select typical hot risk events, enter the online public opinion risk scenario modeling and simulation module, improve the online public opinion risk scenario model, modify and improve the risk level and impact assessment of typical hot events, and form a warning report for the coping strategies proposed by the expert for typical hot risk events.

8. A method for monitoring and warning network public opinion risk based on artificial intelligence, characterized in that, Applied to an online public opinion risk monitoring and warning system based on artificial intelligence as described in any one of claims 1-7, for online public opinion risk monitoring and warning, the method includes: Using the ontology modeling assistance method to construct an online public opinion risk ontology model, and formally expressing the concepts and relationships of the main body, type, propagation law, influencing factors and coping strategies of online public opinion. Based on the online public opinion risk ontology, combined with the data related to online public opinion evaluation obtained from the Internet, construct an online public opinion risk assessment index system; at the same time, preprocess the obtained data, and calculate the measure values of the online public opinion risk assessment indexes of each data in combination with the entropy weight method; predict the future trend of the indexes based on the measure data. Monitor online public opinion risk information, and based on the monitoring information, automatically capture hot risk events in real time, and determine the fields, associated indexes and associated events to which the hot risk events belong. Classify online public opinion risk scenarios, clarify typical risk scenarios, and construct an online public opinion risk scenario model through the system dynamics model; it is also used for simulating the risk scenario model, analyzing the impact of risk events on associated indexes and the impact of coping strategies on online public opinion. Based on the simulation of the risk scenario model, determine the risk level of the associated indexes under the influence of hot risk events, and determine the event risk level through the risk level of the associated indexes. Based on the information of the hot risk event description, the impact on the associated indexes, and the event risk level, use the large language model to generate an evaluation text of the impact of the online public opinion hot risk event. Automatically analyze the field and content of the hot risk event, and generate a warning report for the hot risk event in combination with the evaluation text of the impact of the hot risk event, and send the warning report to the customer terminal in real time. The customer terminal receives the sent warning report and pushes warning information to the customer. Store the online public opinion risk ontology model, the online public opinion risk assessment index system, the online public opinion risk scenario model, the online public opinion typical hot risk events, the online public opinion risk information data and the large language model.

Images