Risk warning and coordinated control methods, systems and equipment based on public opinion analysis

Abstract

This specification discloses a risk warning and linkage control method, system, and device based on public opinion analysis. The method acquires public opinion text data related to a target topic, performs natural language processing on the text data to extract keywords, generates structured semantic feature vectors, inputs these semantic feature vectors into a pre-trained public opinion analysis model, and uses the model to predict public opinion based on the input structured semantic feature vectors, outputting public opinion analysis indicators. These indicators are compared with preset thresholds, and based on the threshold conditions met by the indicators, the public opinion state is analyzed and its trend is predicted, triggering corresponding public opinion state commands. Furthermore, based on these commands, preset linkage control commands related to public opinion changes are executed to adjust the system, introducing real-time public opinion analysis results into the target control chain, thereby enabling timely responses to changes in the public opinion state.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a risk warning and linkage control method, system, and device based on public opinion analysis. Background Technology

[0002] In digital advertising, ad content is often highly tied to brand image. If negative public opinion events occur during the advertising campaign, such as brand controversies, product quality issues, association with social events, or sudden public incidents, and the advertising system continues to run ads according to the original strategy, the negative impact can be easily amplified, and may even trigger a crisis of brand trust.

[0003] Existing advertising systems typically focus on conversion rates, click-through rates, or cost control, lacking sufficient awareness of the external public opinion environment. Public opinion monitoring systems and advertising systems are often independent of each other, resulting in problems such as delayed response, high costs of manual intervention, and coarse-grained control, making it difficult to effectively regulate advertising behavior in the early stages of public opinion risks.

[0004] Therefore, there is an urgent need for a control method that can link with real-time public opinion analysis results, so as to have the ability to proactively perceive public opinion risks and automatically take control measures. Summary of the Invention

[0005] This specification provides a risk warning and linkage control method, system and device based on public opinion analysis to overcome at least one technical problem existing in related technologies.

[0006] According to a first aspect of the embodiments of this specification, a risk warning and linkage control method based on public opinion analysis is provided, including: By using keyword matching, topic subscription, or API calls, public opinion text data related to the target topic is obtained from multiple information sources. Natural language processing is then performed on the public opinion text data to extract text keywords and generate structured semantic feature vectors, which include semantic features of sentiment polarity and sentiment intensity. The semantic feature vector is input into a pre-trained public opinion analysis model. The public opinion analysis model makes predictions based on the input structured semantic feature vector and outputs public opinion analysis indicators. The public opinion analysis model is trained by a neural network model, which filters the semantic feature vector and calculates to predict the trend of public opinion. The public opinion analysis indicators are compared with preset thresholds. Based on the threshold conditions met by the public opinion analysis indicators, the public opinion status is analyzed and the trend of public opinion is predicted. Corresponding public opinion status instructions are triggered. Then, preset linkage control instructions related to changes in public opinion are executed according to the status instructions. The linkage control instructions are associated with the public opinion status instructions and respond to the control target according to changes in the public opinion status.

[0007] Optionally, the steps of obtaining public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription, or interface calls, performing natural language processing on the public opinion text data to extract text keywords, and generating structured semantic feature vectors include: By using keyword matching, topic subscription, or API calls, we can obtain public opinion text data related to the target topic from multiple information sources. Feature extraction is performed on public opinion text data to generate structured semantic feature vectors. , in, Characterizing emotional polarity, Values: -1, 0, 1 Characterizing the intensity of emotion, As a metric for dissemination activity, , Related to the number of users, Related to interaction volume, This represents the platform's influence coefficient. For timestamps.

[0008] Optionally, the step of predicting and outputting public opinion analysis indicators based on the input structured semantic feature vector through a public opinion analysis model includes: Based on the input structured semantic feature vector Calculate public opinion analysis indicators , in, in, The emotional factors are represented by semantic feature vectors that only account for negative emotions. Characterizing the diffusion factor, , , These are the weighting coefficients. , This is the time decay function.

[0009] Optionally, the step of comparing public opinion analysis indicators with preset thresholds, analyzing the public opinion status and predicting the public opinion trend based on the threshold conditions met by the public opinion analysis indicators, and triggering corresponding public opinion status instructions includes: Calculate the rate of change of public opinion indicators based on public opinion analysis indicators. ; Duration parameter for public opinion analysis indicators Perform the counting; The public opinion analysis indicators are compared with a preset first threshold and a second threshold, where the first threshold is less than the second threshold. If the public opinion analysis index is less than the first threshold, the public opinion status is determined to be stable. If the public opinion analysis index is not less than the first threshold and less than the second threshold, the judgment is made based on the sign of the change rate of the public opinion index. If the change rate of the public opinion index is positive, the public opinion status is determined to be in the observation state, and the public opinion observation command is triggered. If the public opinion analysis index is not less than the second threshold and the rate of change of the public opinion index is positive, then the public opinion status is determined to be a risk status, and a public opinion warning instruction is triggered. If the rate of change of public opinion indicators exceeds the third threshold, the public opinion status is determined to be an emergency state, triggering an emergency public opinion command.

[0010] Optionally, based on the results of real-time calculated public opinion analysis indicators and the rate of change of public opinion indicators, combined with the duration parameter of public opinion analysis indicators, when the state is downgraded from an emergency state to a risk state, or from a risk state to an observation state, the corresponding threshold conditions must be met for a preset number of consecutive collection cycles before the state callback instruction is triggered.

[0011] Optionally, after the step of comparing the public opinion analysis indicators with preset first and second thresholds, the method further includes: Calculate the acceleration of change of public opinion indicators based on the rate of change of public opinion indicators. , If the acceleration of change in public opinion indicators exceeds the fourth threshold, the public opinion status is determined to be under observation, triggering a public opinion observation command.

[0012] Optionally, after the step of comparing the public opinion analysis indicators with preset first and second thresholds, the method further includes: The number of state transitions within a preset time window is counted, and the stability constraint parameters are calculated. , in, To calculate the length of the statistical time window, For time window Number of state transitions during the period; When executing state instructions, the stability constraint parameters are... Compared with a preset stability threshold, If stability constraint parameters If the value exceeds the stability threshold, the execution intensity of the currently pending state instruction will be suppressed, delayed, or adjusted.

[0013] According to a second aspect of the embodiments of this specification, a risk warning and linkage control system based on public opinion analysis is provided, including a data acquisition module, a status analysis module, and an instruction output module, wherein... The data acquisition module is configured to acquire public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription or interface call, perform natural language processing on the public opinion text data, extract text keywords, and generate structured semantic feature vectors, the semantic feature vectors including semantic features of sentiment polarity and sentiment intensity. The state analysis module is configured to input the semantic feature vector into a pre-trained public opinion analysis model, and then use the public opinion analysis model to make predictions based on the input structured semantic feature vectors and output public opinion analysis indicators. The public opinion analysis model is trained by a neural network model, and after filtering the semantic feature vectors, it calculates and predicts the trend of public opinion. The instruction output module is configured to compare public opinion analysis indicators with preset thresholds, analyze the public opinion status and predict the public opinion trend based on the threshold conditions met by the public opinion analysis indicators, trigger corresponding public opinion status instructions, and then execute preset linkage control instructions related to public opinion changes according to the status instructions. The linkage control instructions are associated with the public opinion status instructions and respond to the control target according to changes in the public opinion status.

[0014] Optionally, the data acquisition module includes a text acquisition unit and a feature extraction unit, wherein... The text acquisition unit is configured to acquire public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription, or interface calls. The feature extraction unit is configured to extract features from public opinion text data and generate structured semantic feature vectors. , in, Characterizing emotional polarity, Values: -1, 0, 1 Characterizing the intensity of emotion, As a metric for dissemination activity, , Related to the number of users, Related to interaction volume, This represents the platform's influence coefficient. For timestamps.

[0015] According to a third aspect of the embodiments of this specification, a computing device is provided, including a storage device and a processor, wherein the storage device is used to store a computer program, and the processor runs the computer program to cause the computing device to perform the steps of the risk warning and linkage control method based on public opinion analysis.

[0016] The beneficial effects of the embodiments in this specification are as follows: This specification provides a risk warning and linkage control method, system, and device based on public opinion analysis. The method performs natural language processing on public opinion text data, extracts text keywords, generates structured semantic feature vectors, inputs these semantic feature vectors into a pre-trained public opinion analysis model for prediction, outputs public opinion analysis indicators, compares them with preset thresholds, analyzes the public opinion state and predicts its trend based on the comparison results, triggers corresponding public opinion state commands, and then executes preset linkage control commands related to public opinion changes based on the state commands. When applied to an advertising system, after linkage control with the advertising system is executed, it continuously monitors the trend of public opinion changes and the operating status of the advertising system. Based on changes in public opinion, it dynamically adjusts the public opinion risk assessment parameters and linkage control rules, enabling the system to gradually form a more robust and adaptive risk control strategy under different public opinion environments. This achieves real-time perception and rapid response to public opinion risks, reduces the amplification effect of negative brand risks during advertising, improves the operational security of the advertising system, and enhances the adaptability of the advertising system to changes in the external environment.

[0017] The innovative aspects of the embodiments in this specification include: 1. In this specification, continuous data collection is carried out from multiple external information sources. During the collection process, matching and filtering are performed based on keywords and semantic topics. The collected public opinion data is timestamped and source-identified to form a public opinion text sequence. Natural language processing analysis is performed on the public opinion text to extract the sentiment polarity, emotional intensity, and potential risk semantic features of the text. The unstructured public opinion text is transformed into structured semantic analysis results, providing a basic input for subsequent risk quantification. This is one of the innovative points of the embodiments in this specification.

[0018] 2. In this specification, within a preset time window, the semantic analysis results of multiple public opinion texts are aggregated and calculated. Sentiment weight, dissemination activity, and time decay factors are introduced to quantitatively model public opinion risks and obtain a comprehensive risk index reflecting the current brand public opinion status. At the same time, the changes in risk indicators in adjacent time windows are calculated to determine whether the public opinion risk shows an upward trend, thereby avoiding misjudgment due to short-term noise-type negative information. This is one of the innovative points of the embodiments of this specification.

[0019] 3. In this specification, comparing the currently calculated comprehensive public opinion risk indicator and its changing trend with a preset risk judgment rule, when the risk indicator exceeds the corresponding threshold and the risk change rate increases, it is determined that there is a potential advertising risk event with a spreading trend, and an early warning signal is generated. This is one of the innovative points of the embodiments of this specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] To more clearly illustrate the technical solutions in the embodiments of this specification or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0021] Figure 1 It is a schematic flowchart of a risk warning and linkage control method based on public opinion analysis provided by an embodiment of this specification; Figure 2 It is a schematic structural diagram of a risk warning and linkage control system based on public opinion analysis provided by an embodiment of this specification; Figure 3 It is a schematic structural diagram of a computing device provided by an embodiment of this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] The following will clearly and completely describe the technical solutions in the embodiments of this specification in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0023] It should be noted that the terms "include" and "have" and any variations thereof in the embodiments of this specification and the drawings are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0024] The embodiments of this specification disclose a risk warning and linkage control method, system, and device based on public opinion analysis, which will be described in detail below.

[0025] Figure 1 It is a schematic flowchart of a risk warning and linkage control method based on public opinion analysis provided by an embodiment of this specification. As Figure 1As shown, a risk warning and coordinated control method based on public opinion analysis includes: S110. Obtain public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription or interface call, perform natural language processing on the public opinion text data, extract text keywords, and generate structured semantic feature vectors, wherein the semantic feature vectors include semantic features of sentiment polarity and sentiment intensity.

[0026] In a specific embodiment, the steps of S110, which involve obtaining public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription, or interface calls, performing natural language processing on the public opinion text data, extracting text keywords, and generating structured semantic feature vectors, include: By using keyword matching, topic subscription, or API calls, we can obtain public opinion text data related to the target topic from multiple information sources. Feature extraction is performed on public opinion text data to generate structured semantic feature vectors. , in, Characterizing emotional polarity, Values: -1, 0, 1 Characterizing the intensity of emotion, As a metric for dissemination activity, , Related to the number of users, Related to interaction volume, This represents the platform's influence coefficient. For timestamps.

[0027] Continuous data collection is conducted from multiple external information sources related to the brand, product, or content. These sources include at least social media platforms, content comment sections, news websites, and forums. During the collection process, data is matched and filtered based on brand keywords, product names, advertising slogans, or semantic themes. The collected public opinion data is then timestamped and source-identified to form a sequence of public opinion texts that can be used for subsequent analysis.

[0028] S120. Input the semantic feature vector into the pre-trained public opinion analysis model. The public opinion analysis model makes predictions based on the input structured semantic feature vector and outputs public opinion analysis indicators. The public opinion analysis model is trained by a neural network model. After filtering the semantic feature vector, it calculates and predicts the trend of public opinion.

[0029] In a specific embodiment, the step of predicting and outputting public opinion analysis indicators based on the input structured semantic feature vector through a public opinion analysis model includes: Based on the input structured semantic feature vector Calculate public opinion analysis indicators , in, in, The emotional factors are represented by semantic feature vectors that only account for negative emotions. Characterizing the diffusion factor, , , These are the weighting coefficients. , This is the time decay function.

[0030] Within the time window, the sequence of public opinion texts is mapped into a set of structured semantic vectors, and the sentiment polarity value is extracted for each text. For texts with negative sentiment polarity, the risk contribution value is calculated according to the intensity of the negative polarity; for positive or neutral texts, the risk contribution value is zero.

[0031] Meanwhile, a dissemination weight factor based on user influence, interaction activity and platform weight is introduced, and combined with the time decay function, the public opinion risk is weighted and aggregated to obtain the public opinion risk index R(t).

[0032] Public opinion risk indicators are non-negative real numbers, and public opinion analysis indicators are... The smaller the value, the more stable the public opinion situation; public opinion analysis indicators The larger the value, the higher the risk of negative public opinion the brand is currently facing.

[0033] S130. Compare the public opinion analysis indicators with preset thresholds. Based on the threshold conditions met by the public opinion analysis indicators, analyze the public opinion status and predict the public opinion trend. Trigger the corresponding public opinion status instructions. Then, based on the status instructions, execute preset linkage control instructions related to public opinion changes. The linkage control instructions are associated with the public opinion status instructions and respond to the control target according to the changes in public opinion status.

[0034] In a specific embodiment, the steps of comparing public opinion analysis indicators with preset thresholds, analyzing the public opinion status and predicting the public opinion trend based on the threshold conditions met by the public opinion analysis indicators, and triggering corresponding public opinion status commands include: Calculate the rate of change of public opinion indicators based on public opinion analysis indicators. ; Duration parameter for public opinion analysis indicators Perform the counting; The public opinion analysis indicators are compared with a preset first threshold and a second threshold, where the first threshold is less than the second threshold. If the public opinion analysis index is less than the first threshold, the public opinion status is determined to be stable.

[0035] If the public opinion analysis index is not less than the first threshold and is less than the second threshold, the judgment is based on the sign of the rate of change of the public opinion index. If the rate of change of the public opinion index is positive, the public opinion status is determined to be under observation, triggering the public opinion observation command. In the advertising system, this public opinion status is linked to control actions, such as reducing the ad placement intensity by 10%, restricting the placement of high-risk keywords, and increasing the frequency of material review.

[0036] If the public opinion analysis index is not less than the second threshold and the rate of change of the public opinion index is positive, the public opinion status is determined to be a risky state, triggering a public opinion warning command. In the advertising system, this public opinion status is linked to control actions such as reducing resource quotas by 30-50% or restricting high-exposure media.

[0037] If the rate of change of public opinion indicators exceeds the third threshold, the public opinion status is determined to be an emergency, triggering an emergency response command. In the advertising system, this situation corresponds to a sharp increase in risk in a short period of time. The control actions linked to this public opinion status include, for example, suspending all advertising resource configurations, initiating a manual review process, and recording risk logs.

[0038] To prevent frequent status command switching, based on the results of real-time calculated public opinion analysis indicators and the rate of change of public opinion indicators, combined with the duration parameter of public opinion analysis indicators, when switching from an emergency state to a risk state, or from a risk state to an observation state, the status callback command must be triggered only after the corresponding threshold conditions are met for a preset number of consecutive collection cycles.

[0039] In the advertising system, risk status levels can be easily mapped to specific advertising control parameters to achieve automatic control.

[0040] In a specific embodiment, after the step of comparing the public opinion analysis indicators with preset first and second thresholds, the method further includes: Calculate the acceleration of change of public opinion indicators based on the rate of change of public opinion indicators. , If the acceleration of change in public opinion indicators exceeds the fourth threshold, the public opinion status is determined to be under observation, triggering a public opinion observation command.

[0041] The logic for judging the acceleration of changes in public opinion indicators led to the inclusion of an "early warning mechanism" in the methodology. This mechanism not only compares the current risk score with a preset threshold but also comprehensively considers parameters such as the risk change rate and the duration of the risk, constructing a segmented risk control state model. Simultaneously, the control state of the linkage control system is divided into normal, observation, risk, and emergency states, and switching between these states is based on the risk value range and its changing trend.

[0042] In a specific embodiment, after the step of comparing the public opinion analysis indicators with preset first and second thresholds, the method further includes: The number of state transitions within a preset time window is counted, and the stability constraint parameters are calculated. , in, To calculate the length of the statistical time window, For time window Number of state transitions during the period; When executing state instructions, the stability constraint parameters are... Compared with a preset stability threshold, If stability constraint parameters If the value exceeds the stability threshold, the execution intensity of the currently pending state instruction will be suppressed, delayed, or adjusted.

[0043] By introducing stability constraint parameters, the repeated triggering of state command switching within the risk critical range is avoided, thus increasing stability.

[0044] Figure 2 This is a schematic diagram of a risk warning and linkage control system based on public opinion analysis, provided as an embodiment of this specification. Figure 2 As shown, a risk warning and linkage control system 200 based on public opinion analysis includes a data acquisition module 210, a status analysis module 220, and an instruction output module 230, wherein... The data acquisition module 210 is configured to acquire public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription or interface call, perform natural language processing on the public opinion text data, extract text keywords, and generate a structured semantic feature vector, which includes semantic features of sentiment polarity and sentiment intensity.

[0045] The state analysis module 220 is configured to input the semantic feature vector into a pre-trained public opinion analysis model, and then use the public opinion analysis model to make predictions based on the input structured semantic feature vectors and output public opinion analysis indicators. The public opinion analysis model is trained by a neural network model, and after filtering the semantic feature vectors, it calculates and predicts the trend of public opinion.

[0046] The instruction output module 230 is configured to compare public opinion analysis indicators with preset thresholds, analyze the public opinion status and predict the public opinion trend based on the threshold conditions met by the public opinion analysis indicators, trigger corresponding public opinion status instructions, and then execute preset linkage control instructions related to public opinion changes according to the status instructions. The linkage control instructions are associated with the public opinion status instructions and respond to the control target according to the changes in the public opinion status.

[0047] In a specific embodiment, the data acquisition module includes a text acquisition unit and a feature extraction unit, wherein... The text acquisition unit is configured to acquire public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription, or interface calls.

[0048] The feature extraction unit is configured to extract features from public opinion text data and generate structured semantic feature vectors. , in, Characterizing emotional polarity, Values: -1, 0, 1 Characterizing the intensity of emotion, As a metric for dissemination activity, , Related to the number of users, Related to interaction volume, This represents the platform's influence coefficient. For timestamps.

[0049] Figure 3 This is a schematic diagram of the structure of a computing device provided in one embodiment of this specification. Figure 3 As shown, a computing device 300 includes a storage device 310 and a processor 320. The storage device 310 is used to store computer programs, and the processor 320 runs the computer programs to enable the computing device 300 to perform the steps of the risk warning and linkage control method based on public opinion analysis.

[0050] In summary, the embodiments of this specification provide a risk warning and linkage control method, system and device based on public opinion analysis. This method constructs a linkage mechanism between public opinion perception, risk assessment and target control system, and automatically intervenes in the linkage target control behavior when public opinion risk occurs or is about to occur, thereby reducing risk and improving the security and robustness of the target control system.

[0051] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of one embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing the present invention.

[0052] Those skilled in the art will understand that the modules in the apparatus of the embodiments can be distributed in the apparatus of the embodiments as described in the embodiments, or they can be located in one or more devices different from this embodiment with corresponding changes. The modules of the above embodiments can be combined into one module, or they can be further divided into multiple sub-modules.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A risk warning and coordinated control method based on public opinion analysis, characterized in that, include: By using keyword matching, topic subscription, or API calls, public opinion text data related to the target topic is obtained from multiple information sources. Natural language processing is then performed on the public opinion text data to extract text keywords and generate structured semantic feature vectors, which include semantic features of sentiment polarity and sentiment intensity. The semantic feature vector is input into a pre-trained public opinion analysis model. The public opinion analysis model makes predictions based on the input structured semantic feature vector and outputs public opinion analysis indicators. The public opinion analysis model is trained by a neural network model, which filters the semantic feature vector and calculates to predict the trend of public opinion. The public opinion analysis indicators are compared with preset thresholds. Based on the threshold conditions met by the public opinion analysis indicators, the public opinion status is analyzed and the trend of public opinion is predicted. Corresponding public opinion status instructions are triggered. Then, preset linkage control instructions related to changes in public opinion are executed according to the status instructions. The linkage control instructions are associated with the public opinion status instructions and respond to the control target according to changes in the public opinion status.

2. The method according to claim 1, characterized in that, The steps described above, including obtaining public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription, or API calls, performing natural language processing on the public opinion text data to extract text keywords, and generating structured semantic feature vectors, include: By using keyword matching, topic subscription, or API calls, we can obtain public opinion text data related to the target topic from multiple information sources. Feature extraction is performed on public opinion text data to generate structured semantic feature vectors. , in, Characterizing emotional polarity, Values: -1, 0, 1 Characterizing the intensity of emotion, As a metric for dissemination activity, , Related to the number of users, Related to interaction volume, This represents the platform's influence coefficient. For timestamps.

3. The method according to claim 2, characterized in that, The steps described above for predicting and outputting public opinion analysis indicators based on the input structured semantic feature vector through a public opinion analysis model include: Based on the input structured semantic feature vector Calculate public opinion analysis indicators , in, in, The emotional factors are represented by semantic feature vectors that only account for negative emotions. Characterizing the diffusion factor, , , These are the weighting coefficients. , This is the time decay function.

4. The method according to claim 1, characterized in that, The steps of comparing public opinion analysis indicators with preset thresholds, analyzing the public opinion status and predicting the public opinion trend based on the threshold conditions met by the public opinion analysis indicators, and triggering corresponding public opinion status instructions include: Calculate the rate of change of public opinion indicators based on public opinion analysis indicators. ; Duration parameter for public opinion analysis indicators Perform the counting; The public opinion analysis indicators are compared with a preset first threshold and a second threshold, where the first threshold is less than the second threshold. If the public opinion analysis index is less than the first threshold, the public opinion status is determined to be stable. If the public opinion analysis index is not less than the first threshold and less than the second threshold, the judgment is made based on the sign of the change rate of the public opinion index. If the change rate of the public opinion index is positive, the public opinion status is determined to be in the observation state, and the public opinion observation command is triggered. If the public opinion analysis index is not less than the second threshold and the rate of change of the public opinion index is positive, then the public opinion status is determined to be a risk status, and a public opinion warning instruction is triggered. If the rate of change of public opinion indicators exceeds the third threshold, the public opinion status is determined to be an emergency state, triggering an emergency public opinion command.

5. The method according to claim 4, characterized in that, Based on the real-time calculated public opinion analysis indicators and the rate of change of public opinion indicators, combined with the duration parameter of public opinion analysis indicators, when the state is downgraded from an emergency state to a risk state, or from a risk state to an observation state, the corresponding threshold conditions must be met for a preset number of consecutive collection cycles before the state callback instruction is triggered.

6. The method according to claim 4, characterized in that, Following the step of comparing the public opinion analysis indicators with preset first and second thresholds, the method further includes: Calculate the acceleration of change of public opinion indicators based on the rate of change of public opinion indicators. , If the acceleration of change in public opinion indicators exceeds the fourth threshold, the public opinion status is determined to be under observation, triggering a public opinion observation command.

7. The method according to claim 4, characterized in that, Following the step of comparing the public opinion analysis indicators with preset first and second thresholds, the method further includes: The number of state transitions within a preset time window is counted, and the stability constraint parameters are calculated. , in, To calculate the length of the statistical time window, For time window Number of state transitions during the period; When executing state instructions, the stability constraint parameters are... Compared with a preset stability threshold, If stability constraint parameters If the value exceeds the stability threshold, the execution intensity of the currently pending state instruction will be suppressed, delayed, or adjusted.

8. A risk warning and linkage control system based on public opinion analysis, characterized in that, It includes a data acquisition module, a status analysis module, and an instruction output module, among which... The data acquisition module is configured to acquire public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription or interface call, perform natural language processing on the public opinion text data, extract text keywords, and generate structured semantic feature vectors, the semantic feature vectors including semantic features of sentiment polarity and sentiment intensity. The state analysis module is configured to input the semantic feature vector into a pre-trained public opinion analysis model, and then use the public opinion analysis model to make predictions based on the input structured semantic feature vectors and output public opinion analysis indicators. The public opinion analysis model is trained by a neural network model, and after filtering the semantic feature vectors, it calculates and predicts the trend of public opinion. The instruction output module is configured to compare public opinion analysis indicators with preset thresholds, analyze the public opinion status and predict the public opinion trend based on the threshold conditions met by the public opinion analysis indicators, trigger corresponding public opinion status instructions, and then execute preset linkage control instructions related to public opinion changes according to the status instructions. The linkage control instructions are associated with the public opinion status instructions and respond to the control target according to changes in the public opinion status.

9. The system according to claim 8, characterized in that, The data acquisition module includes a text acquisition unit and a feature extraction unit, wherein... The text acquisition unit is configured to acquire public opinion text data related to the target topic from multiple information sources through keyword matching, topic subscription, or interface calls. The feature extraction unit is configured to extract features from public opinion text data and generate structured semantic feature vectors. , in, Characterizing emotional polarity, Values: -1, 0, 1 Characterizing the intensity of emotion, As a metric for dissemination activity, , Related to the number of users, Related to interaction volume, This represents the platform's influence coefficient. For timestamps.

10. A computing device, characterized in that, The device includes a storage device and a processor, the storage device being used to store a computer program, and the processor running the computer program to cause the computing device to perform the steps of the method according to any one of claims 1-7.

Images