Internet public opinion intelligent intervention system and method based on big data

Abstract

The invention discloses an intelligent intervention system and method for network public opinions based on big data. The intelligent intervention system comprises a data acquisition module which is mainly used for acquiring text, image and video data in real time; the data processing module is mainly used for performing targeted processing on text, image and video data; the public opinion feature extraction module is mainly used for constructing a public opinion propagation rate curve and extracting emotion features; the public opinion propagation prediction model module is mainly used for capturing time sequence relevance between features through a bidirectional LSTM network by adopting a deep neural network architecture, and introducing an attention mechanism to dynamically allocate feature weights; and the public opinion processing module is mainly used for constructing an intervention strategy matrix according to the propagation trend predicted value and the emotional evolution path output by the prediction model so as to realize differential processing of public opinions with different risk levels. The method has the advantages that through the bidirectional LSTM and the attention mechanism, according to different risk levels, the time sequence relevance and emotion changes of public opinion propagation are accurately captured, and real-time control and accurate response are achieved.

Description

Technical Field

[0001] The present invention relates to big data processing, and particularly to an intelligent intervention system and method for online public opinion based on big data. Background Art

[0002] With the rapid development of social media, news platforms, and self-media, the information dissemination mode has undergone a profound transformation, showing significant characteristics of fragmentation, instantaneity, and omnipresence. In terms of fragmented dissemination, short contents such as short videos, microblogs, and social dynamics have become the mainstream. Information is disassembled into more easily spread but possibly distorted fragments, resulting in a rapid shift of public opinion foci and a one-sided public perception. The instantaneity trend is manifested as the "zero time difference" between information release and dissemination. An emergency can spread across the entire network within minutes through forms such as live broadcasts and news flashes, and the public opinion fermentation cycle has been shortened from several days in the traditional era to several hours or even shorter. Omnipresence is reflected in the popularization of the mobile Internet, which enables information to penetrate into every corner of society. Anyone can become an information publisher and disseminator through a smart phone, breaking the information monopoly pattern in the traditional media era.

[0003] Most of the current online public opinion intervention methods rely on traditional rules and keyword matching, with limited processing speed and accuracy, and lack of in-depth understanding and prediction of the public opinion dissemination process. Many methods can often only deal with the public opinion crises that have already occurred, and the intervention means are relatively single, lacking flexibility and real-time performance. Traditional methods mainly rely on manual analysis and empirical judgment, which easily lead to lagged processing of public opinion evolution and cannot effectively predict the evolution trend and emotional changes of public opinion. In addition, these methods have not fully utilized deep learning and sentiment analysis technologies, cannot accurately mine the temporal correlation in the public opinion dissemination process, and lack comprehensive processing and integration of multi-modal data. This makes the existing methods lack flexible response capabilities and dynamic adjustment strategies in the face of complex and changeable public opinion environments, often resulting in the failure to timely and effectively intervene in public opinion, which may trigger a greater range of social impacts. Summary of the Invention

[0004] In order to improve the existing online public opinion intervention system and method, an intelligent intervention system and method for online public opinion based on big data are provided. This method combines bidirectional LSTM and attention mechanism to achieve in-depth mining of the temporal correlation of public opinion dissemination and dynamic risk assessment, and constructs personalized and adaptive intelligent intervention strategies based on multi-modal data fusion and sentiment analysis.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows: An intelligent intervention system for online public opinion based on big data, comprising: A data acquisition module: The data acquisition module is mainly used for real-time acquisition of text, image, and video data from online platforms such as social media, news websites, and forums; Data processing module: The data processing module is mainly used to perform targeted processing on text, image, and video data respectively; Public opinion feature extraction module: The public opinion feature extraction module is mainly used to construct a public opinion propagation rate curve through time series features and extract emotional features from the collected data through an emotional dictionary; Public opinion propagation prediction model module: The public opinion propagation prediction model module is mainly used to adopt a deep neural network architecture, capture the temporal correlation between features through a bidirectional LSTM network, and introduce an attention mechanism to dynamically allocate feature weights; Public opinion processing module: The public opinion processing module is mainly used to construct an intervention strategy matrix based on the propagation trend prediction value and emotional evolution path output by the prediction model, and realize the differential processing of public opinions with different risk levels; Database module: The database module is mainly used to store the collected data, model training data, and historical processing records.

[0006] Preferably, the data processing module specifically includes: Text processing unit: The text processing unit performs deduplication processing, non-text data stripping, sensitive information filtering, and word segmentation processing on text data; Image processing unit: The image processing unit performs metadata extraction, sensitive information filtering, and image scene recognition processing on image data; Video processing unit: The video processing unit performs key frame extraction, visual detection processing, and audio processing on video data; Based on the data collected by each unit, a structured public opinion database is constructed by aggregation.

[0007] Preferably, the public opinion feature extraction module specifically includes: Time series feature construction unit: Based on the public opinion propagation speed data in the structured public opinion database, construct a public opinion propagation speed curve model; Calculate the first and second derivatives based on the public opinion propagation speed curve, obtain the inflection points of the curve, and define the outbreak inflection point, decay inflection point, and secondary propagation inflection point; Generate and obtain the features of public opinion changing with time by designing a decay function, and construct a feature vector; Emotional feature extraction unit: Based on the data in the structured public opinion database, match it with an open-source dictionary that integrates HowNet and BosonNLP to obtain data containing emotions; Based on the obtained data containing emotions, set the division weights according to the emotional intensity, and dynamically adjust the weights based on the time-dependent decay; Based on the dissemination chain of public opinion data, conduct dissemination level division, construct an emotional dissemination tree, extract horizontal and vertical features from the tree structure, and classify emotional dissemination patterns, including: explosive type, controversial type, attenuation type; Feature fusion unit: The feature fusion unit is mainly used to splice the time series feature vector and the emotional tree feature into a joint feature, and obtain key indicators, including: emotional mutation detection during the outbreak period, emotional consistency index during the attenuation period.

[0008] Preferably, the public opinion dissemination prediction model module specifically includes: Bidirectional LSTM unit: The bidirectional LSTM unit is mainly used to obtain the time series correlation between joint features through a bidirectional LSTM network model based on the fused joint feature data; Attention mechanism unit: The attention mechanism unit is mainly used to dynamically assign weights to joint features according to their importance, and obtain key factors similar to public opinion dissemination; Training unit: The training unit is mainly used to train the model through a bidirectional LSTM network model based on the historical public opinion data in the database module, and obtain a trained public opinion dissemination prediction model.

[0009] Preferably, the public opinion processing module specifically includes: Public opinion risk assessment unit: The public opinion risk assessment unit is mainly used to input the real-time collected public opinion data into the public opinion dissemination prediction model, obtain the public opinion dissemination speed change curve and public opinion emotion evolution path in the future time period, and evaluate the public opinion risk level based on the real-time collected data and the prediction model data; Control strategy unit: The control strategy unit is mainly used to design an intervention strategy matrix according to the evaluated risk level, design personalized control strategies for the public opinion changes in various risk index situations, and construct a strategy dynamic execution system for multi-platform collaborative control and adversarial disposal.

[0010] Furthermore, a method for intelligent intervention of big data network public opinion includes: Real-time collect multi-modal data such as text, images, and videos through network platforms such as social media, news websites, and forums; Based on the obtained data, perform deduplication, format standardization, sensitive information filtering, and structured processing to construct a structured public opinion database; Based on the structured public opinion database, obtain the time series features and emotional features of public opinion dissemination, splice and fuse the time series features and emotional features to construct a joint feature vector; Based on a deep learning model of bidirectional LSTM and attention mechanism, train and predict the public opinion dissemination trend and emotional evolution path, and construct a public opinion dissemination prediction model; Based on the prediction results, obtain the propagation speed and sentiment trend of public opinion, evaluate its risk level, and construct a public opinion intervention strategy matrix according to different risk levels to achieve precise classification management and real-time control and response to different public opinion risks.

[0011] Compared with the prior art, the advantages of the present invention are as follows: Through multi-modal data collection, refined data processing, and deep learning modeling, it can capture and analyze the dynamics of online public opinion in real time. Its greatest advantage lies in being able to quickly identify the propagation trend and sentiment changes of public opinion, providing precise prediction support for public opinion management. By combining a bidirectional LSTM network with an attention mechanism, it can deeply mine the temporal correlation of public opinion propagation and formulate personalized intervention strategies for each type of risk. In addition, through emotion feature extraction and risk assessment, the system can dynamically adjust the strategy to ensure that intervention measures are quickly and effectively implemented in different public opinion environments, avoiding the spread of negative emotions. By combining multi-platform collaborative control, it provides comprehensive technical support for public opinion monitoring and crisis management, and is a highly intelligent and flexible and controllable public opinion intervention solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a schematic diagram of the system proposed by the present invention; Figure 2 It is a schematic diagram of the method proposed by the present invention; Figure 3 It is a diagram of the data processing module proposed by the present invention; Figure 4 It is a diagram of the public opinion feature extraction module proposed by the present invention; Figure 5 It is a diagram of the public opinion propagation prediction model module proposed by the present invention; Figure 6 It is a diagram of the public opinion processing module proposed by the present invention; Figure 7 It is an architecture diagram of the electronic device in this solution; Figure 8 It is a schematic diagram of the structure of the computer-readable storage medium in this solution. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments in the following description are only examples, and those skilled in the art can think of other obvious variations.

[0014] Refer to Figure 1 As shown, an intelligent intervention system for online public opinion based on big data includes: Data collection module: The data collection module is mainly used to collect text, image, and video data of network platforms such as social media, news websites, and forums in real time; Data processing module: The data processing module is mainly used to perform targeted processing on text, image, and video data respectively; Public opinion feature extraction module: The public opinion feature extraction module is mainly used to construct a public opinion propagation rate curve through temporal features and extract emotional features from the collected data through an emotional dictionary; Public opinion propagation prediction model module: The public opinion propagation prediction model module is mainly used to adopt a deep neural network architecture, capture the temporal correlation between features through a bidirectional LSTM network, and introduce an attention mechanism to dynamically allocate feature weights; Public opinion processing module: The public opinion processing module is mainly used to construct an intervention strategy matrix according to the predicted propagation trend value and emotional evolution path output by the prediction model, and realize the differential processing of public opinions with different risk levels; Database module: The database module is mainly used to store the collected data, model training data, and historical processing records.

[0015] Refer to Figure 2 As shown, an intelligent intervention method for online public opinion based on big data includes: Real-time collection of multi-modal data such as text, images, and videos through online platforms such as social media, news websites, and forums; Based on the acquired data, perform deduplication, format normalization, sensitive information filtering, and structured processing to construct a structured public opinion database; Based on the structured public opinion database, obtain the temporal features and emotional features of public opinion propagation, splice and fuse the temporal features and emotional features, and construct a joint feature vector; Based on a deep learning model of bidirectional LSTM and attention mechanism, train and predict the public opinion propagation trend and emotional evolution path, and construct a public opinion propagation prediction model; Based on the prediction results, obtain the propagation speed and emotional trend of public opinion, evaluate its risk level, and construct a public opinion intervention strategy matrix according to different risk levels to achieve precise classification management and real-time control and response to different public opinion risks.

[0016] Refer to Figure 3 As shown, the data processing module specifically includes: Text processing unit: The text processing unit performs deduplication processing, non-text data stripping, sensitive information filtering, and word segmentation on text data; Image processing unit: The image processing unit extracts metadata, filters sensitive information, and performs image scene recognition on image data; Video processing unit: The video processing unit extracts key frames, performs visual detection processing, and audio processing on video data; Based on the data collected by each unit, gather and construct a structured public opinion database.

[0017] Specifically, when processing text data, first convert the text data into a word sequence and segment the text through the database dictionary data; Based on the data after this segmentation, eliminate duplicate or highly similar text data, and generate text fingerprints through the SimHash algorithm for text feature extraction. The formula is:

[0018] where, is the word item after word segmentation, is the word hash value, is the word weight; Based on the characteristics of each text word item, calculate the similarity between each word item. The formula is:

[0019] Based on the threshold, determine whether each word item belongs to similar text and merge the duplicate items; After that, delete the non-text data in the text data. Through regular expressions or pattern matching techniques, identify and remove the non-text data. Common non-text data types include URLs, HTML tags, and image links; Based on the sensitive information dataset in the database, first perform fuzzy matching to obtain a general range, and then perform exact matching within the determined range. By traversing the sensitive word library, check whether the text contains sensitive words; When processing image data, use an image processing library to extract metadata from the image file, identify faces, license plates, etc. in the image through a deep learning model, and identify and filter sensitive content from the text in the image through text recognition technology; When processing video data, extract the most representative frames from the video stream and use the inter-frame difference method to detect significant changes. The formula is:

[0020] where, is the pixel value (grayscale) of the t-th frame, is the threshold; If , then the current frame is a candidate key frame. Remove redundancy from the candidate frames and calculate the similarity between two frames through structural similarity. The formula is:

[0021] where, is the mean value, is the standard deviation, , are constants; For the audio data in the video data, the spectrum is extracted through short-time Fourier transform, and the voiceprint is recognized.

[0022] Refer to Figure 4 As shown, the public opinion feature extraction module specifically includes: Time series feature construction unit: Based on the public opinion dissemination speed data in the structured public opinion database, construct a public opinion dissemination speed curve model; Calculate the first and second derivatives based on the public opinion dissemination speed curve, obtain the inflection points of the curve, and define the outbreak inflection point, decay inflection point, and secondary dissemination inflection point; By designing a decay function, generate and obtain the features of public opinion changing over time, and construct a feature vector; Emotional feature extraction unit: Based on the data in the structured public opinion database, match it with an open-source dictionary that integrates HowNet and BosonNLP to obtain data containing emotions; Based on the obtained data containing emotions, set the division weights according to the emotional intensity, and dynamically adjust the weights based on timeliness decay; Based on the public opinion data dissemination chain, conduct dissemination level division, construct an emotional dissemination tree, and extract horizontal and vertical features from the tree structure, and classify the emotional dissemination modes, including: explosion type, controversy type, decay type; Feature fusion unit: The feature fusion unit is mainly used to splice the time series feature vector and the emotional tree feature into a joint feature, and obtain key indicators, including: emotional mutation detection during the outbreak period, emotional consistency index during the decay period.

[0023] Specifically, extract the hourly dissemination volume of public opinion from the database, perform time series aggregation, and smooth the noise through a Gaussian kernel. The formula is:

[0024] where V(t) is the dissemination volume, =3, k = 3 ; Fit the dissemination speed through the Logistic-Growth model. The formula is:

[0025] where, is the maximum dissemination volume, is the dissemination rate, is the inflection point time; By taking the first and second derivatives of the public opinion dissemination speed curve, obtain the inflection points of the curve, and classify the inflection point types, including: Outbreak inflection point: ; Decay inflection point: ; Secondary propagation inflection point: ; The exponential decay correction model is designed and adopted to characterize the propagation decay stage, and the formula is:

[0026] where is the initial decay amount, is the decay rate, is the steady-state baseline; Based on the above characteristics, a time series feature vector is constructed to represent the public opinion life cycle; When constructing the emotional propagation tree, the user forwarding relationship graph , the node corresponds to the text , and , and the one with the maximum emotional intensity is selected as the root of the propagation tree. When dividing the levels, the propagation depth is defined as the shortest path length to the root, ; For each propagation tree, a feature vector is generated, and a three-class SVM is used. The kernel function is selected as RBF, and the emotional propagation mode is divided into explosive, controversial, and decay types. The kernel function formula is:

[0027] where is the feature vector.

[0028] Refer to Figure 5 As shown, the public opinion propagation prediction model module specifically includes: Bidirectional LSTM unit: The bidirectional LSTM unit is mainly used to obtain the temporal correlation between joint features through the bidirectional LSTM network model based on the fused joint feature data; Attention mechanism unit: The attention mechanism unit is mainly used to dynamically assign weights to the joint features according to their importance to obtain the key factors similar to public opinion propagation; Training unit: The training unit is mainly used to train the model through the bidirectional LSTM network model based on the historical public opinion data in the database module to obtain the trained public opinion propagation prediction model.

[0029] Specifically, by fusing emotional, propagation structure, and timeliness features, a time series feature matrix is constructed:

[0030] where is the time step, is the feature dimension; The temporal correlation between joint features is obtained through bidirectional LSTM forward propagation. The forward LSTM unit calculation formula is: , the reverse LSTM calculation formula is: , the bidirectional state splicing formula is: ; The attention score of each feature vector is calculated based on the time step importance, and dynamic weight allocation is performed to train the model.

[0031] See Figure 6 As shown in the figure, the public opinion processing module specifically includes: Public opinion risk assessment unit: The public opinion risk assessment unit is mainly used to input the public opinion data collected in real time into the public opinion propagation prediction model, obtain the public opinion propagation speed change curve and public opinion sentiment evolution path in the future time period, and evaluate the public opinion risk level based on the real-time collected data and the prediction model data; Control strategy unit: The control strategy unit is mainly used to design an intervention strategy matrix based on the assessed risk level, design personalized control strategies based on changes in public opinion under various risk indicators, and build a dynamic strategy execution system to conduct multi-platform collaborative control and adversarial disposal.

[0032] Specifically, various risk indicators are quantified, as follows:

[0033] Based on the risk indicators after beam formation, dynamic risk grading rules are set, specifically: Red risk (immediate action): Spread potential > 80% and emotional polarization > 0.6; exponential spread of more than two derivative topics detected; Orange risk (key monitoring): Transmission potential 60%-80% or key node rate >30%; cross-platform index exceeds the baseline value by 3σ; Blue risk (routine observation): Spread potential <30% and no derivative topics; The risk value is recalculated every 30 minutes using the time decay correction factor. When the risk degrades for two consecutive periods, the alarm is triggered and lifted. The formula is:

[0034] Based on the assessed risk level, an intervention strategy matrix is designed to divide the strategy dimensions:

[0035] Combination based on multimodal strategies, specifically:

[0036] Further, the method according to the embodiments of the present application can also be implemented by means of Figure 7 the architecture of the electronic device shown. As Figure 7 shown, the electronic device 500 may include a bus 501, one or more CPUs 502, a read-only memory (ROM) 503, a random access memory (RAM) 504, a communication port 505 connected to a network, an input / output component 506, a hard disk 507, etc. The storage device in the electronic device 500, such as the ROM 503 or the hard disk 507, can store a big data network public opinion intelligent intervention system and method provided by the present application. The electronic device 500 may further include a terminal interface 508. Of course, Figure 7 the architecture shown is only exemplary. When implementing different devices, one or more components in the Figure 7 shown electronic device can be omitted according to actual needs.

[0037] Figure 8 FIG. is a schematic structural diagram of a computer-readable storage medium provided by an embodiment of the present application. As Figure 8 shown, it is a computer-readable storage medium 600 according to an embodiment of the present application. Computer-readable instructions are stored on the computer-readable storage medium 600. When the computer-readable instructions are run by a processor, a big data network public opinion intelligent intervention system and method according to the embodiments of the present application described with reference to the above drawings can be executed. The storage medium 600 includes but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory, etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0038] It should be noted that: the above sequence of the embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. Additionally, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0039] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments.

[0040] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. An intelligent intervention system for big data network public opinion, characterized in that, Including: Data acquisition module: The data acquisition module is mainly used to collect text, image, and video data from online platforms such as social media, news websites, and forums in real time; Data processing module: The data processing module is mainly used to perform targeted processing on text, image, and video data respectively; Public opinion feature extraction module: The public opinion feature extraction module is mainly used to construct a public opinion propagation rate curve through time series features and extract emotional features from the collected data using an emotion dictionary; Public opinion propagation prediction model module: The public opinion propagation prediction model module is mainly used to adopt a deep neural network architecture, capture the temporal correlation between features through a bidirectional LSTM network, and introduce an attention mechanism to dynamically allocate feature weights; Public opinion processing module: The public opinion processing module is mainly used to construct an intervention strategy matrix based on the predicted propagation trend value and emotional evolution path output by the prediction model, and realize the differential processing of public opinions with different risk levels; Database module: The database module is mainly used to store the collected data, model training data, and historical processing records.

2. The intelligent intervention system for big data network public opinion according to claim 1, wherein The data processing module specifically includes: Text processing unit: The text processing unit performs deduplication processing, non-text data stripping, sensitive information filtering, and word segmentation on text data; Image processing unit: The image processing unit extracts metadata, filters sensitive information, and performs image scene recognition on image data; Video processing unit: The video processing unit extracts key frames, performs visual detection processing, and audio processing on video data; Based on the data collected by each unit, a structured public opinion database is constructed by aggregation.

3. The intelligent intervention system for online public opinion based on big data according to claim 1, wherein, The public opinion feature extraction module specifically includes: Time series feature construction unit: Based on the public opinion propagation speed data in the structured public opinion database, a public opinion propagation speed curve model is constructed; Calculate the first-order and second-order derivatives based on the public opinion propagation speed curve, obtain the inflection points of the curve, and define the outbreak inflection point, decay inflection point, and secondary propagation inflection point; Generate and obtain the features of public opinion changing over time by designing a decay function, and construct a feature vector; Emotional feature extraction unit: Based on the data in the structured public opinion database, match it with an open-source dictionary that integrates HowNet and BosonNLP to obtain data containing emotions; Based on the obtained data containing emotions, set the division weights according to the emotional intensity, and dynamically adjust the weights based on the timeliness decay; Based on the public opinion data propagation chain, conduct propagation level division, construct an emotional propagation tree, and perform horizontal and vertical feature extraction on the tree structure, and classify the emotional propagation modes, including: explosion type, controversy type, decay type; Feature fusion unit: The feature fusion unit is mainly used to splice the time series feature vector and the emotional tree feature into a joint feature, and obtain key indicators, including: emotional mutation detection during the outbreak period, emotional consistency index during the decay period.

4. An intelligent intervention system for online public opinion based on big data according to claim 1, characterized in that, The public opinion propagation prediction model module specifically includes: Bidirectional LSTM unit: The bidirectional LSTM unit is mainly used to obtain the temporal correlation between joint features through a bidirectional LSTM network model based on the fused joint feature data; Attention mechanism unit: The attention mechanism unit is mainly used to dynamically allocate weights to the joint features according to their importance, and obtain the key factors for public opinion dissemination. Training unit: The training unit is mainly used to train the model through a bidirectional LSTM network model based on the historical public opinion data in the database module, and obtain a trained public opinion dissemination prediction model.

5. An intelligent intervention system for big data network public opinion according to claim 1, characterized in that, The public opinion processing module specifically includes: Public opinion risk assessment unit: The public opinion risk assessment unit is mainly used to input the real-time collected public opinion data into the public opinion dissemination prediction model, obtain the change curve of the public opinion dissemination speed and the emotional evolution path of the public opinion in the future time period, and evaluate the public opinion risk level based on the real-time collected data and the prediction model data. Control strategy unit: The control strategy unit is mainly used to design an intervention strategy matrix according to the evaluated risk level, design personalized control strategies for the public opinion changes under various risk indicators, and construct a dynamic strategy execution system for multi-platform collaborative control and adversarial disposal.

6. An intelligent intervention method for online public opinion based on big data, characterized in that, Including: Real-time collection of multi-modal data such as text, images, and videos through network platforms such as social media, news websites, and forums; Based on the obtained data, perform deduplication, format standardization, sensitive information filtering, and structured processing to construct a structured public opinion database; Based on the structured public opinion database, obtain the temporal characteristics and emotional characteristics of public opinion dissemination, splice and fuse the temporal characteristics and emotional characteristics to construct a joint feature vector; Based on a deep learning model of bidirectional LSTM and attention mechanism, train and predict the public opinion dissemination trend and emotional evolution path, and construct a public opinion dissemination prediction model; Based on the prediction results, obtain the dissemination speed and emotional trend of public opinion, evaluate its risk level, and construct a public opinion intervention strategy matrix according to different risk levels to achieve precise classification management and real-time control and response to different public opinion risks.

7. An electronic device, characterized in that, Including: At least one processor; And a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a method for intelligent intervention of big data network public opinion as described in any one of claims 1-6.

8. A computer-readable storage medium storing computer-readable instructions, characterized in that, When the computer-readable instructions are executed by the processor, a method for intelligent intervention of big data network public opinion as described in any one of claims 1-6 is implemented.

Images