Multi-dimensional rumor detection and prevention system based on large language model

Abstract

The invention discloses a multi-dimensional rumor detection and prevention system based on a large language model, which relates to the technical field of rumor detection and comprises a data collection module, a data preprocessing module, a feature extraction module, a background knowledge integration module, a logical reasoning module, a detection and classification module and a prevention and response module. According to the rumor detection method and system, all aspects of the rumor can be comprehensively considered from multiple angles, the aspects include but are not limited to the content of the rumor, the propagation path of the rumor and the interaction condition between the users, and through the comprehensive analysis, the system can ensure that detection of the rumor covers all related dimensions, so that the accuracy and comprehensiveness of detection are improved, and the detection efficiency is improved. The rumor detection and classification work can be efficiently carried out by utilizing an advanced artificial intelligence technology, and through an intelligent means, the system not only improves the detection efficiency, but also remarkably improves the detection accuracy and the intelligent level, so that the rumor detection is more accurate and efficient.

Description

Technical Field

[0001] The present invention relates to the technical field of rumor detection, and in particular to a multi-dimensional rumor detection and prevention system based on a large language model. Background Art

[0002] Rumor detection is of great significance for maintaining social order, protecting public interests, and promoting the true dissemination of information. First, it helps prevent the spread of false information and avoid causing public panic or misleading decision-making. Secondly, by promptly identifying and correcting rumors, it can maintain the reputation of individuals and institutions and reduce unnecessary economic losses and social conflicts. In addition, rumor detection can also help enhance the sense of responsibility of the media and social platforms and promote a healthy information ecology. Finally, it also plays a positive role in enhancing the public's information literacy and critical thinking ability.

[0003] The existing technology has the following problems:

[0004] Traditional rumor detection algorithms usually only rely on the text content of the rumor and related information during its propagation process, but often ignore the background knowledge and logical relationships behind the rumor. This single detection method leads to low comprehensiveness and intelligence, and is unable to effectively deal with the complex and changeable rumor propagation phenomenon. In addition, these algorithms lack the ability to dynamically update and cannot adapt to new rumor patterns and propagation methods in a timely manner, making them incapable of doing so in practical applications. Summary of the invention

[0005] The present invention provides a multi-dimensional rumor detection and prevention system based on a large language model to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] The multi-dimensional rumor detection and prevention system based on a large language model includes a data collection module, a data preprocessing module, a feature extraction module, a background knowledge integration module, a logic reasoning module, a detection and classification module, and a prevention and response module. The data collection module integrates social media crawlers, API interfaces, and news aggregators. The data preprocessing module covers data cleaning, word segmentation and annotation, encoding conversion, and other links. The feature extraction module focuses on extracting key features such as semantics, emotions, and style. The background knowledge integration module uses knowledge graphs and fact databases to build a rich knowledge base. The logic reasoning module uses strategies such as contradiction detection and reasoning verification. The detection and classification module uses machine learning algorithms and deep learning models to achieve accurate classification. The prevention and response module effectively responds to rumors through measures such as automatic labeling, user reminders, and content deletion.

[0008] This system relies on the following core technologies to achieve its functions:

[0009] Fine-tuning of large language models: By using labeled rumor and non-rumor datasets, the pre-trained large language models are fine-tuned to enable them to accurately capture the deceptiveness and information inconsistency of rumors, thereby improving their performance in rumor detection tasks.

[0010] Knowledge graph construction: With the help of entity recognition technology, the system can accurately extract key information from rumor texts, laying a solid foundation for building a knowledge graph. Subsequently, by determining the complex relationships between entities, the content of the knowledge graph is further enriched, providing the system with more in-depth analysis and reasoning capabilities;

[0011] Logical reasoning algorithm: Through consistency checking rules, source credibility rules, time logic rules, context relevance rules, and reasoning strength rules, the system can analyze and understand the logical structure and factual relationship in rumor texts;

[0012] Classification algorithm design: The system uses advanced algorithms of decision trees and random forests to accurately identify rumor types. The decision tree can be expanded into a multi-label classifier by applying multiple label segmentation criteria to each node, while the random forest is used as an integration method to integrate the results of multiple decision trees to improve prediction accuracy, thereby enabling the system to more flexibly respond to various rumor challenges and significantly improve classification accuracy.

[0013] A further improvement of the technical solution of the present invention is that the social media crawler uses an advanced distributed technology architecture to achieve efficient crawling of cross-platform data and has excellent high-concurrency processing capabilities. This crawler system can automatically capture posts, comments and user information on mainstream social media platforms such as Weibo, WeChat, and Douyin, ensuring the timeliness and richness of the data. The API interface establishes a stable cooperative relationship with major social media platforms through security authentication mechanisms such as OAuth, and safely and stably accesses their API interfaces, thereby obtaining data streams in real time, further improving the timeliness and accuracy of the data. The news aggregator makes full use of diversified means such as RSS subscriptions and API calls to achieve instant updating of news data. The aggregator widely covers multiple information sources such as news websites, blogs, forums, etc., ensuring the comprehensiveness and depth of rumor detection.

[0014] A further improvement of the technical solution of the present invention is that the data cleaning process adopts regular expressions and natural language processing tools to achieve accurate elimination of invalid data, including the removal of noise data such as HTML tags, URLs, etc., thereby ensuring high standards of data quality; the word segmentation and annotation link uses cutting-edge word segmentation algorithms and part-of-speech tagging tools, such as BERT, spaCy, etc., to perform detailed word segmentation on the text and annotate part-of-speech, entity and other information, laying a solid foundation for subsequent data processing; the encoding conversion function supports multiple character encodings, such as UTF-8, GBK, etc., ensuring data compatibility, achieving the unification of data formats, greatly facilitating subsequent processing flows, and improving the overall compatibility of the system.

[0015] A further improvement of the technical solution of the present invention is that: the semantic features use large language models such as BERT and GPT to deeply explore the deep semantic information of the text. These models can extract the deep semantic features of the text, thereby revealing the true meaning of the text. The emotional features use sentiment analysis models such as TextBl ob and VADER to evaluate the emotional tendency of the text, such as positive, negative, neutral, etc., to help identify potential rumors. The style features combine machine learning and deep learning methods to accurately identify the style features of the text, such as formal, informal, humorous, etc., further enriching the detection dimension.

[0016] A further improvement of the technical solution of the present invention is that the knowledge graph is constructed and maintained using advanced graph databases such as Neo4j. The carefully constructed graph integrates multiple elements such as entities, relationships and attributes, providing a solid and rich background knowledge support for the logical reasoning process. The fact database ensures the timeliness and cutting-edge nature of the collected information through a regular update mechanism. The database is specifically used to integrate strictly verified factual data and news reports to ensure the accuracy of the detection results.

[0017] A further improvement of the technical solution of the present invention is that the contradiction detection mechanism will integrate natural language reasoning technology to accurately identify the logical contradictions hidden in the text. This mechanism is specifically used to analyze the content of rumors, reveal the logical contradictions and inconsistencies between the facts, and thus effectively reveal the false nature of the rumors. The reasoning verification process cleverly utilizes the rich resources of the knowledge graph and the fact database, and not only performs reasoning based on established rules, but also uses model-driven methods for in-depth analysis. In this process, the background knowledge base becomes an indispensable support, assisting each step of the reasoning verification and ensuring the accuracy and reliability of the final detection results.

[0018] The further improvement of the technical solution of the present invention lies in that: the machine learning algorithm uses classic methods such as SVM and random forest, focuses on rumor classification, and significantly enhances the accuracy of rumor detection by integrating classification algorithms. The deep learning model uses convolutional neural networks and recurrent neural networks to further optimize the accuracy of rumor detection. These deep learning models not only deepen the level of rumor detection, but also improve the intelligence level of the detection process.

[0019] Further improvements to the technical solution of the present invention are: the automatic marking system will be committed to developing an efficient automatic marking function to quickly identify and mark rumor content, ensure that these rumors can be automatically marked, and facilitate subsequent processing; the user reminder module will integrate a push notification service to send rumor warnings to users in a timely manner, thereby enhancing users' awareness of prevention and ensuring that users can quickly alert and respond to rumors; the content deletion mechanism will actively cooperate with social media platforms to jointly implement automatic or semi-automatic content deletion processes, thereby effectively curbing the spread of rumors and maintaining the health and stability of the network environment.

[0020] Due to the adoption of the above technical solution, the present invention has the following technical advances compared with the prior art:

[0021] 1. The present invention provides a multi-dimensional rumor detection and prevention system based on a large language model. The system is comprehensive and can comprehensively consider various aspects of rumors from multiple angles, including but not limited to the content of the rumor itself, its propagation path, and the interaction between users. Through this comprehensive analysis, the system can ensure that the detection of rumors covers all relevant dimensions, thereby improving the accuracy and comprehensiveness of detection.

[0022] 2. The present invention provides a multi-dimensional rumor detection and prevention system based on a large language model. The system also has intelligent characteristics. It uses advanced artificial intelligence technology to efficiently perform rumor detection and classification. Through intelligent means, the system not only improves the detection efficiency, but also significantly improves the detection accuracy and intelligence level, making rumor detection more accurate and efficient.

[0023] 3. The present invention provides a multi-dimensional rumor detection and prevention system based on a large language model. The system also has the ability to be dynamically updated. As new data and information continue to emerge, the system can continue to learn and self-update to adapt to the ever-changing rumor propagation environment. This dynamic update mechanism ensures the timeliness of the rumor detection system, enabling it to respond to newly emerging rumors in a timely manner, thereby maintaining the continuity and effectiveness of the detection effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a system architecture diagram of the multi-dimensional rumor detection and prevention system based on a large language model of the present invention;

[0025] Figure 2 This is an operation flow chart of the multi-dimensional rumor detection and prevention system based on a large language model of the present invention. DETAILED DESCRIPTION

[0026] The present invention is further described in detail below in conjunction with embodiments:

[0027] Example 1

[0028] like Figure 1-2 As shown, the present invention provides a multi-dimensional rumor detection and prevention system based on a large language model, including a data collection module, a data preprocessing module, a feature extraction module, a background knowledge integration module, a logical reasoning module, a detection and classification module, and a prevention and response module. The data collection module integrates social media crawlers, API interfaces, and news aggregators. The data preprocessing module covers data cleaning, word segmentation and annotation, encoding conversion, and other links. The feature extraction module focuses on extracting key features such as semantic features, emotional features, and style features. The background knowledge integration module uses knowledge graphs and fact databases to build a rich knowledge base. The logical reasoning module uses strategies such as contradiction detection and reasoning verification. The detection and classification module uses machine learning algorithms and deep learning models to achieve accurate classification. The prevention and response module effectively responds to rumors through measures such as automatic labeling, user reminders, and content deletion.

[0029] This system relies on the following core technologies to achieve its functions:

[0030] Fine-tuning of large language models: By using labeled rumor and non-rumor datasets, the pre-trained large language models are fine-tuned to enable them to accurately capture the deceptiveness and information inconsistency of rumors, thereby improving their performance in rumor detection tasks.

[0031] Knowledge graph construction: With the help of entity recognition technology, the system can accurately extract key information from rumor texts, laying a solid foundation for building a knowledge graph. Subsequently, by determining the complex relationships between entities, the content of the knowledge graph is further enriched, providing the system with more in-depth analysis and reasoning capabilities;

[0032] Logical reasoning algorithm: Through consistency checking rules, source credibility rules, time logic rules, context relevance rules, and reasoning strength rules, the system can analyze and understand the logical structure and factual relationship in rumor texts;

[0033] Classification algorithm design: The system uses advanced algorithms of decision trees and random forests to accurately identify rumor types. The decision tree can be expanded into a multi-label classifier by applying multiple label segmentation criteria to each node, while the random forest is used as an integration method to integrate the results of multiple decision trees to improve prediction accuracy, thereby enabling the system to more flexibly respond to various rumor challenges and significantly improve classification accuracy.

[0034] In this embodiment, the system aims to effectively identify and respond to online rumors through a series of highly specialized modules and core technologies. The system consists of modules such as data collection, preprocessing, feature extraction, background knowledge integration, logical reasoning, detection and classification, and prevention and response, covering the entire process from data collection to rumor processing. The core technologies include fine-tuning of large language models, construction of knowledge graphs, and design of logical reasoning algorithms and classification algorithms. By fine-tuning the large language model to capture rumor features, building a knowledge graph to enhance analysis capabilities, applying logical reasoning algorithms to understand rumor logic, and adopting advanced classification algorithms to improve recognition accuracy, the system can flexibly respond to a variety of rumors, and excel in rumor detection and prevention, providing strong support for the health and security of the network environment.

[0035] Example 2

[0036] like Figure 1-2 As shown, on the basis of Example 1, the present invention provides a technical solution: Preferably, the social media crawler uses an advanced distributed technology architecture to achieve efficient crawling of cross-platform data and has excellent high-concurrency processing capabilities. This crawler system can automatically capture posts, comments and user information on mainstream social media platforms such as Weibo, WeChat, and Douyin, ensuring the timeliness and richness of the data. The API interface establishes a stable cooperative relationship with major social media platforms through security authentication mechanisms such as OAuth, and safely and stably accesses their API interfaces, thereby obtaining data streams in real time, further improving the timeliness and accuracy of the data. The news aggregator makes full use of diversified means such as RSS subscriptions and API calls to achieve instant updating of news data. The aggregator widely covers multiple information sources such as news websites, blogs, forums, etc., ensuring the comprehensiveness and depth of rumor detection.

[0037] The data cleaning process uses regular expressions and natural language processing tools to accurately eliminate invalid data, including removing noise data such as HTML tags, URLs, etc., so as to ensure high standards of data quality. The word segmentation and annotation process uses cutting-edge word segmentation algorithms and part-of-speech tagging tools, such as BERT, spaCy, etc., to perform detailed word segmentation on the text and annotate part-of-speech, entity and other information, laying a solid foundation for subsequent data processing. The encoding conversion function supports multiple character encodings, such as UTF-8, GBK, etc., ensuring data compatibility, realizing the unification of data formats, greatly facilitating subsequent processing flows, and improving the overall compatibility of the system.

[0038] The semantic features use large language models such as BERT and GPT to deeply explore the deep semantic information of the text. These models can extract the deep semantic features of the text, thereby revealing the true meaning of the text. The emotional features use sentiment analysis models such as TextBl ob and VADER to evaluate the emotional tendency of the text, such as positive, negative, neutral, etc., to help identify potential rumors. The style features combine machine learning and deep learning methods to accurately identify the style features of the text, such as formal, informal, humorous, etc., further enriching the detection dimension.

[0039] The knowledge graph is constructed and maintained using advanced graph databases such as Neo4j. The carefully constructed graph integrates multiple elements such as entities, relationships and attributes, providing a solid and rich background knowledge support for the logical reasoning process. The fact database ensures the timeliness and cutting-edge nature of the collected information through a regular update mechanism. The database is specifically used to integrate strictly verified factual data and news reports to ensure the accuracy of the test results.

[0040] The contradiction detection mechanism will integrate natural language reasoning technology to accurately identify potential logical contradictions in the text. This mechanism is specifically used to analyze the content of rumors, reveal the logical contradictions and inconsistencies between the facts, and thus effectively reveal the false nature of the rumors. The reasoning verification process cleverly utilizes the rich resources of the knowledge graph and fact database, and not only performs reasoning based on established rules, but also uses model-driven methods for in-depth analysis. In this process, the background knowledge base becomes an indispensable support, assisting each step of the reasoning verification and ensuring the accuracy and reliability of the final detection results.

[0041] The machine learning algorithm uses classic methods such as SVM and random forest, focuses on rumor classification, and significantly enhances the accuracy of rumor detection by integrating classification algorithms. The deep learning model uses convolutional neural networks and recurrent neural networks to further optimize the accuracy of rumor detection. These deep learning models not only deepen the level of rumor detection, but also improve the intelligence level of the detection process.

[0042] The automatic tagging system will be committed to developing efficient automatic tagging functions to quickly identify and tag rumor content, ensure that these rumors can be automatically marked, and facilitate subsequent processing. The user reminder module will integrate push notification services to send rumor warnings to users in a timely manner, thereby enhancing users' awareness of prevention and ensuring that users can quickly alert and respond to rumors. The content deletion mechanism will actively cooperate with social media platforms to jointly implement automatic or semi-automatic content deletion processes, thereby effectively curbing the spread of rumors and maintaining the health and stability of the network environment.

[0043] In this embodiment, the system achieves efficient data capture through a distributed technical architecture and has high concurrent processing capabilities. The system can automatically capture posts, comments and user information from platforms such as Weibo, WeChat, and Douyin, and cooperate with social media platforms through a security authentication mechanism to obtain data streams in real time. The news aggregator uses RSS subscriptions and API calls to achieve instant updates of news data. The data cleaning process uses regular expressions and natural language processing tools to eliminate invalid data and ensure data quality. The word segmentation and annotation links use cutting-edge algorithms to lay the foundation for data processing. The encoding conversion function supports multiple character encodings to improve system compatibility. The extraction of semantic features, emotional features, and style features utilizes large language models and analysis tools. The knowledge graph and fact database provide background knowledge support for logical reasoning. The contradiction detection mechanism and the reasoning verification process use knowledge graphs and fact databases to reveal the false nature of rumors. Machine learning and deep learning models optimize the accuracy of rumor detection. The automatic marking system, user reminder module, and content deletion mechanism work together to effectively curb the spread of rumors and maintain a healthy and stable network environment.

[0044] The following is a detailed explanation of the working principle of this multi-dimensional rumor detection and prevention system based on a large language model.

[0045] like Figure 1-2As shown in the figure, first of all, the system uses the data collection module to automatically capture posts, comments and user information on social media platforms, and at the same time works closely with social media platforms to obtain data in real time with the help of API interfaces to ensure comprehensive coverage and instant updates of data. In addition, the system also widely collects content from news websites and blogs to achieve comprehensive coverage of rumor detection. Subsequently, the data preprocessing module carefully cleans the collected data and effectively removes noise data, such as HTML tags, URLs, etc., to ensure high quality and purity of the data. The module also further performs word segmentation on the text and marks parts of speech and entities, laying a solid foundation for subsequent feature extraction and processing. At the same time, encoding conversion is performed to ensure the uniformity of the data format and enhance the compatibility and stability of the system. In the feature extraction module, the system uses advanced large-scale language models to deeply explore the deep semantic features of the text, analyze emotional tendencies, and accurately identify the style of the text, thereby accurately revealing the truth of the text. The background knowledge integration module constructs a knowledge graph containing entities, relationships and attributes, integrating verified facts and news reports, providing rich and reliable background knowledge support for logical reasoning, and ensuring the accuracy of the detection results. The logical reasoning module deeply analyzes the logical contradictions and factual inconsistencies in the rumors, and uses the background knowledge base for rigorous reasoning and verification, which effectively reveals the false nature of the rumors and ensures the reliability and authority of the detection results. The detection and classification module uses advanced classification algorithms and deep learning models to accurately classify and efficiently detect rumors, significantly improving the accuracy and intelligence of detection. Finally, the prevention and response module automatically marks the detected rumor content for subsequent processing and tracking. At the same time, the system also sends rumor warning information to users in a timely manner to enhance users' awareness of prevention and self-protection capabilities. In addition, the system also works closely with the platform to jointly delete false information and effectively curb the spread and diffusion of rumors.

[0046] The above generally describes the present invention in detail, but it is obvious to a person skilled in the art that some modifications or improvements can be made to the present invention. Therefore, modifications or improvements that do not depart from the spirit of the present invention are within the scope of protection of the present invention.

Claims

1. A multi-dimensional rumor detection and prevention system based on a large language model, characterized by: It includes a data collection module, a data preprocessing module, a feature extraction module, a background knowledge integration module, a logical reasoning module, a detection and classification module, and a prevention and response module. The data collection module integrates social media crawlers, API interfaces, and news aggregators. The data preprocessing module covers data cleaning, word segmentation and annotation, encoding conversion, and other links. The feature extraction module focuses on extracting key features such as semantic features, emotional features, and style features. The background knowledge integration module uses knowledge graphs and fact databases to build a rich knowledge base. The logical reasoning module uses strategies such as contradiction detection and reasoning verification. The detection and classification module uses machine learning algorithms and deep learning models to achieve accurate classification. The prevention and response module effectively responds to rumors through measures such as automatic labeling, user reminders, and content deletion. This system relies on the following core technologies to achieve its functions: Fine-tuning of large language models: By using labeled rumor and non-rumor datasets, the pre-trained large language models are fine-tuned to enable them to accurately capture the deceptiveness and information inconsistency of rumors, thereby improving their performance in rumor detection tasks. Knowledge graph construction: With the help of entity recognition technology, the system can accurately extract key information from rumor texts, laying a solid foundation for building a knowledge graph. Subsequently, by determining the complex relationships between entities, the content of the knowledge graph is further enriched, providing the system with more in-depth analysis and reasoning capabilities; Logical reasoning algorithm: Through consistency checking rules, source credibility rules, time logic rules, context relevance rules, and reasoning strength rules, the system can analyze and understand the logical structure and factual relationship in rumor texts; Classification algorithm design: The system uses advanced algorithms of decision trees and random forests to accurately identify rumor types. The decision tree can be expanded into a multi-label classifier by applying multiple label segmentation criteria to each node, while the random forest is used as an integration method to integrate the results of multiple decision trees to improve prediction accuracy, thereby enabling the system to more flexibly respond to various rumor challenges and significantly improve classification accuracy.

2. The multi-dimensional rumor detection and prevention system based on a large language model according to claim 1, characterized in that: The social media crawler uses an advanced distributed technology architecture to achieve efficient crawling of cross-platform data and has excellent high-concurrency processing capabilities. This crawler system can automatically capture posts, comments and user information on mainstream social media platforms such as Weibo, WeChat, and Douyin, ensuring the timeliness and richness of the data. The API interface establishes a stable cooperative relationship with major social media platforms through security authentication mechanisms such as OAuth, and safely and stably accesses their API interfaces to obtain data streams in real time, further improving the timeliness and accuracy of the data. The news aggregator makes full use of diversified means such as RSS subscriptions and API calls to achieve instant updating of news data. The aggregator covers a wide range of information sources such as news websites, blogs, forums, etc., ensuring the comprehensiveness and in-depth nature of rumor detection.

3. The multi-dimensional rumor detection and prevention system based on a large language model according to claim 1 is characterized in that: The data cleaning process uses regular expressions and natural language processing tools to accurately eliminate invalid data, including removing noise data such as HTML tags, URLs, etc., so as to ensure high standards of data quality. The word segmentation and annotation process uses cutting-edge word segmentation algorithms and part-of-speech tagging tools, such as BERT, spaCy, etc., to perform detailed word segmentation on the text and annotate part-of-speech, entity and other information, laying a solid foundation for subsequent data processing. The encoding conversion function supports multiple character encodings, such as UTF-8, GBK, etc., ensuring data compatibility, realizing the unification of data formats, greatly facilitating subsequent processing flows, and improving the overall compatibility of the system.

4. The multi-dimensional rumor detection and prevention system based on a large language model according to claim 1, characterized in that: The semantic features use large language models such as BERT and GPT to deeply explore the deep semantic information of the text. These models can extract the deep semantic features of the text, thereby revealing the true meaning of the text. The emotional features use sentiment analysis models such as TextBlob and VADER to evaluate the emotional tendency of the text, such as positive, negative, neutral, etc., to help identify potential rumors. The style features combine machine learning and deep learning methods to accurately identify the style features of the text, such as formal, informal, humorous, etc., further enriching the detection dimension.

5. The multi-dimensional rumor detection and prevention system based on a large language model according to claim 1, characterized in that: The knowledge graph is constructed and maintained using advanced graph databases such as Neo4j. The carefully constructed graph integrates multiple elements such as entities, relationships and attributes, providing a solid and rich background knowledge support for the logical reasoning process. The fact database ensures the timeliness and cutting-edge nature of the collected information through a regular update mechanism. The database is specifically used to integrate strictly verified factual data and news reports to ensure the accuracy of the test results.

6. The multi-dimensional rumor detection and prevention system based on a large language model according to claim 1, characterized in that: The contradiction detection mechanism will integrate natural language reasoning technology to accurately identify potential logical contradictions in the text. This mechanism is specifically used to analyze the content of rumors, reveal the logical contradictions and inconsistencies between the facts, and thus effectively reveal the false nature of the rumors. The reasoning verification process cleverly utilizes the rich resources of the knowledge graph and fact database, and not only performs reasoning based on established rules, but also uses model-driven methods for in-depth analysis. In this process, the background knowledge base becomes an indispensable support, assisting each step of the reasoning verification and ensuring the accuracy and reliability of the final detection results.

7. The multi-dimensional rumor detection and prevention system based on a large language model according to claim 1, characterized in that: The machine learning algorithm uses classic methods such as SVM and random forest, focuses on rumor classification, and significantly enhances the accuracy of rumor detection by integrating classification algorithms. The deep learning model uses convolutional neural networks and recurrent neural networks to further optimize the accuracy of rumor detection. These deep learning models not only deepen the level of rumor detection, but also improve the intelligence level of the detection process.

8. The multi-dimensional rumor detection and prevention system based on a large language model according to claim 1, characterized in that: The automatic tagging system will be committed to developing efficient automatic tagging functions to quickly identify and tag rumor content, ensure that these rumors can be automatically marked, and facilitate subsequent processing. The user reminder module will integrate push notification services to send rumor warnings to users in a timely manner, thereby enhancing users' awareness of prevention and ensuring that users can quickly alert and respond to rumors. The content deletion mechanism will actively cooperate with social media platforms to jointly implement automatic or semi-automatic content deletion processes, thereby effectively curbing the spread of rumors and maintaining the health and stability of the network environment.

Images