Cross-platform information propagation analysis method, system, equipment and product

Abstract

The invention belongs to the technical field of network information processing, and aims to provide a cross-platform information propagation analysis method, system, equipment and product. The method comprises the following steps: performing multi-modal data acquisition processing from different platforms to obtain multi-source data, and constructing a knowledge graph according to the multi-source data; performing event topic mining processing and sentiment analysis processing on the knowledge graph to obtain event topic distribution information and a sentiment analysis result, and constructing a propagation graph according to the event topic distribution information and the sentiment analysis result; and obtaining an entity type set, an inter-entity relationship and an entity attribute according to the propagation atlas, and constructing a path tracing database according to the entity type set, the inter-entity relationship and the entity attribute so as to carry out cross-platform information propagation tracing based on the path tracing database. According to the method, cross-platform information propagation analysis can be realized, and the propagation path analysis capability is high.

Description

Technical Field

[0001] This invention belongs to the field of network information processing technology, specifically relating to a cross-platform information dissemination analysis method, system, device, and product. Background Technology

[0002] In recent years, with the rapid development of Internet technology, various online social media platforms have emerged. In order to help governments, enterprises and research institutions accurately grasp public opinion dynamics, optimize marketing strategies, and identify false information, so as to improve the efficiency and scientific nature of information governance, it is necessary to conduct dissemination analysis of information on the Internet.

[0003] However, in the digital age, information dissemination has broken through the limitations of single platforms, forming a complex network that is cross-platform (such as social media, news websites, short video platforms, and forums) and multimodal (including text, images, video, and audio). Simultaneously, the speed and scope of information dissemination have significantly increased. This has led the inventors to discover, in using existing technologies, that traditional information dissemination analysis methods suffer from at least the following problems: First, traditional methods can only process data from a single platform or a single modality, making it difficult to integrate heterogeneous data from different platforms and modalities. For example, it is difficult to correlate and analyze text and image information on social media platforms with bullet comments on short video platforms, resulting in low information utilization. Second, traditional methods rely on simple statistics or fixed rules, making it difficult to capture variations such as content rewriting, splicing, or secondary creation during the dissemination process, resulting in distorted dissemination paths, insufficient accuracy in tracing sources, and weak dissemination path analysis capabilities. Summary of the Invention

[0004] The present invention aims to solve the above-mentioned technical problems to at least a certain extent, and provides a cross-platform information dissemination analysis method, system, device and product.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a cross-platform information dissemination analysis method, comprising: Multimodal data is collected and processed from different platforms to obtain multi-source data, and a knowledge graph is constructed based on the multi-source data. The knowledge graph is subjected to event topic mining and sentiment analysis to obtain event topic distribution information and sentiment analysis results, and a propagation graph is constructed based on the event topic distribution information and sentiment analysis results. The propagation graph is used to obtain the entity type set, the relationship between entities and the entity attributes. A path tracing database is constructed based on the entity type set, the relationship between entities and the entity attributes so as to carry out cross-platform information propagation tracing based on the path tracing database.

[0006] In one possible design, a knowledge graph is constructed based on the multi-source data, including: The multi-source data is subjected to structured parsing to obtain a structured data pool; Cross-modal feature fusion processing is performed on the media resource data in the structured data pool to obtain a joint embedding space; Entities, relationships, and attributes are extracted based on the joint embedding space, and a knowledge graph is constructed based on the entities, relationships, and attributes.

[0007] In one possible design, after forming the joint embedding space, the method further includes: A stratified sampling method is used to perform stratified sampling on the embedding representations in the joint embedding space to obtain sampled data. The quality of the sampled data is then evaluated, and if the quality evaluation fails, the embedding representations in the joint embedding space are completed to obtain a corrected joint embedding space, so as to construct a knowledge graph based on the corrected joint embedding space.

[0008] In one possible design, the knowledge graph is subjected to event topic mining and sentiment analysis to obtain event topic distribution information and sentiment analysis results, including: The BERTopic topic model is used to perform semantic clustering on the content entities in the knowledge graph to extract event topic distribution information; wherein, the event topic distribution information includes the event topic described by each content entity; The DeBERTa-v3 pre-trained model is used to perform text sentiment analysis on the content entities in the knowledge graph to obtain sentiment analysis results, which include sentiment type and intensity.

[0009] In one possible design, the entity type set, inter-entity relationships, and entity attributes are obtained based on the propagation graph, including: The propagation map is used to identify entity boundaries to obtain a set of entity types. The implicit relationships between entities in the propagation graph are learned using a graph attention network, and the implicit relationships between entities are verified and completed using a business rule base to obtain the final relationships between entities. Cross-validation and conflict resolution are performed on the entity attributes in the propagation graph to obtain the final entity attributes.

[0010] In one possible design, the path tracing database is stored in RDF format and cross-platform information dissemination and tracing are performed through the SPARQL query interface.

[0011] In one possible design, after constructing the propagation graph, the method further includes: Based on the propagation map, a multidimensional evaluation index is constructed, and each evaluation index is normalized to obtain a comprehensive evaluation value of the propagation index.

[0012] Secondly, the present invention provides a cross-platform information dissemination analysis system, comprising: The data fusion module is used to collect and process multimodal data from different platforms to obtain multi-source data, and to construct a knowledge graph based on the multi-source data. The propagation graph construction module is communicatively connected to the data fusion module and is used to perform event topic mining and sentiment analysis on the knowledge graph to obtain event topic distribution information and sentiment analysis results, and construct a propagation graph based on the event topic distribution information and sentiment analysis results. The propagation tracing module is communicatively connected to the propagation graph construction module. It is used to obtain the entity type set, the relationship between entities and the entity attributes based on the propagation graph, and to construct a path tracing database based on the entity type set, the relationship between entities and the entity attributes, so as to perform cross-platform information propagation tracing based on the path tracing database.

[0013] Thirdly, the present invention provides an electronic device, comprising: Memory, used to store computer program instructions; and, A processor is configured to execute the computer program instructions to perform the operation of a cross-platform information dissemination analysis method as described in any of the preceding claims.

[0014] Fourthly, the present invention provides a computer program product, including a computer program or instructions, wherein the computer program or instructions, when executed by a computer, implement a cross-platform information propagation analysis method as described in any one of the above.

[0015] The beneficial effects of this invention are as follows: This invention discloses a cross-platform information dissemination analysis method, system, device, and product, which can realize cross-platform information dissemination analysis and has strong dissemination path analysis capabilities. Specifically, in the implementation process, this invention integrates heterogeneous information from cross-platform multimodal data to construct a dynamic dissemination map based on user interaction behavior and content characteristics, and realizes dissemination path tracing. This can break through the data silo limitations of traditional methods, and significantly improve the real-time performance and accuracy of cross-platform dissemination analysis, enhancing the dissemination path analysis capabilities.

[0016] Other beneficial effects of the present invention will be further explained in the specific embodiments. Attached Figure Description

[0017] Figure 1 This is a flowchart of the cross-platform information propagation analysis method in Example 1; Figure 2 This is a block diagram of the cross-platform information dissemination and analysis system in Example 2; Figure 3 This is a block diagram of an electronic device in Embodiment 3. Detailed Implementation

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0019] Example 1: This embodiment discloses a cross-platform information dissemination analysis method, which can be executed by, but is not limited to, computer devices or virtual machines with certain computing resources, such as personal computers, smartphones, personal digital assistants or wearable devices, or by virtual machines.

[0020] like Figure 1 As shown, a cross-platform information dissemination analysis method may include, but is not limited to, the following steps: S1. Multimodal data is collected and processed from different platforms to obtain multi-source data, and a knowledge graph is constructed based on the multi-source data. Specifically, in this embodiment, the multi-source data includes text, images, videos, and audio data collected from different platforms. During implementation, a multi-source data collection engine is pre-deployed, and data crawling interfaces are customized for the characteristics of different platforms such as social media, short video platforms, and news websites to ensure comprehensive collection of full-modal data such as text, images, videos, and audio.

[0021] In step S1, constructing a knowledge graph based on the multi-source data includes: S101. Perform structured parsing on the multi-source data to obtain a structured data pool. Specifically, in this embodiment, a structured processing pipeline is used to convert unstructured data in the multi-source data into structured metadata. During implementation, a heterogeneous data processing pipeline is pre-built, integrating an OCR (Optical Character Recognition) recognition module, a YOLOv8 (an efficient target detection algorithm based on a single-stage architecture) target detection module, and an ASR (Automatic Speech Recognition) speech-to-text module, etc., to perform structured parsing on unstructured data such as text, audio, images, and videos, thereby obtaining a structured data pool containing metadata such as keyframes, object labels, and voiceprint features.

[0022] S102. Perform cross-modal feature fusion processing on the media resource data in the structured data pool to obtain a joint embedding space. Specifically, in this embodiment, the joint embedding space can also be called a text-image-video joint embedding space. In the implementation process, firstly, the media resource data in the structured data pool is encoded to output embedding vectors of different modalities. For example, a pre-trained BERT (Bidirectional Encoder Representations from Transformers) model is used to encode text segments in the structured data pool, outputting a 768-dimensional text embedding vector. The CLIP-ViT-L / 14 model is used to encode keyframe images in the structured data pool, outputting a 1024-dimensional image embedding vector. This is not limited here. Subsequently, the InfoNCE (Information Noise-Contrastive Estimation) loss function is used to modally align the embedding vectors of text and image modalities to obtain a joint embedding representation. Finally, the joint embedding representations are concatenated and then subjected to overall dimensionality reduction to obtain the joint embedding space.

[0023] Furthermore, in this embodiment, not only can media resource information such as text and images be associated, but user behavior information generated by propagation in the structured data pool can also be integrated. During this process, by associating user and media resource usage data, Graph Attention Networks (GATs) are constructed, thereby enabling structured display in the subsequent knowledge graph with users as entity nodes and user content consumption patterns as propagation edges.

[0024] S103. The embedding representations in the joint embedding space are subjected to stratified sampling using a stratified sampling method to obtain sampling data. The quality of the sampling data is then evaluated, and if the quality evaluation fails, the embedding representations in the joint embedding space are completed to obtain a corrected joint embedding space, so as to construct a knowledge graph based on the corrected joint embedding space.

[0025] Specifically, in this embodiment, firstly, a stratified sampling method is used to stratify the embedding representations in the joint embedding space according to media type (text / image / video), with 5% of the data extracted from each layer. User behavior information is clustered and sampled by device ID to ensure coverage of cross-platform behavior. Subsequently, the sampled data is checked to ensure completeness by checking whether key fields are empty, to ensure rationality by checking whether the text embedding vector is within the reasonable range of the pre-trained BERT model, and to ensure timeliness by calculating the difference between the data generation time and the current time. Finally, missing or abnormal data that do not meet the quality assessment requirements are generatively completed by integrating KNN-Imputer and GAN, thereby obtaining a corrected joint embedding space for constructing a knowledge graph.

[0026] S104. Entities, relationships, and attributes are extracted based on the joint embedding space, and a knowledge graph is constructed based on the entities, relationships, and attributes. Specifically, in this embodiment, firstly, the BERT-BiLSTM-CRF (Bidirectional Encoder Representations from Transformers-Bidirectional Long Short-Term Memory-Conditional Random Field) model is used to perform entity recognition processing on the joint embedding space to extract entities such as people / places / organizations from the joint embedding space; then, a "precursor-follower" propagation relationship chain is constructed through features such as event time sequence; subsequently, attributes on the propagation path are supplemented, such as the "forward count" statistics of the cross-platform propagation path; finally, a knowledge graph is constructed based on the entities, relationships, and attributes.

[0027] S2. Perform event topic mining and sentiment analysis on the knowledge graph to obtain event topic distribution information and sentiment analysis results, and construct a propagation graph based on the event topic distribution information and sentiment analysis results; specifically, in this embodiment, the propagation graph uses users or content as nodes, and uses event topic distribution information and sentiment analysis results as additional attributes of the nodes, and calculates the weight of the propagation edges based on the behavior type between users and content in the knowledge graph (such as forwarding, commenting, liking), time decay factor and platform weight, to construct a dynamic propagation graph containing node attributes and weighted directed edges, thereby realizing a structured representation of the cross-platform information propagation process.

[0028] In step S2, the knowledge graph is subjected to event topic mining and sentiment analysis to obtain event topic distribution information and sentiment analysis results, including: S201. The BERTopic topic model is used to perform semantic clustering on the content entities in the knowledge graph to extract event topic distribution information; wherein, the event topic distribution information includes the event topics described by each content entity; in this embodiment, by using the BERTopic topic model to perform semantic clustering on the content entities, the core topic distribution of the event is automatically identified, revealing the topic categories to which different content belongs, and enhancing the understanding of the event structure.

[0029] S202. Use the DeBERTa-v3 pre-trained model to perform text sentiment analysis on the content entities in the knowledge graph to obtain sentiment analysis results; wherein, the sentiment analysis results include sentiment type and intensity. The sentiment type adopts fine-grained sentiment type, including positive, negative, and neutral, etc. By accurately capturing public sentiment tendencies, data support can be provided for public opinion analysis, risk warning, etc.

[0030] In this embodiment, by performing event topic mining and sentiment analysis on the knowledge graph, the semantic core and emotional tendency of the dissemination content can be extracted from the content entities in the knowledge graph. This provides a key data foundation for the subsequent construction of a dissemination graph with semantic and emotional attributes, which helps to enhance the accuracy and interpretability of dissemination analysis.

[0031] S3. Obtain the entity type set, entity relationships, and entity attributes based on the propagation graph, and construct a path tracing database based on the entity type set, entity relationships, and entity attributes to facilitate cross-platform information propagation tracing. It should be noted that the path tracing database is used to systematically store entities, events, and their relationships during the information propagation process, supporting subsequent propagation path queries and tracing analysis.

[0032] In step S3, the entity type set, inter-entity relationships, and entity attributes are obtained based on the propagation graph, including: S301. Entity boundary identification is performed on the propagation graph using the BERT-BiLSTM-CRF model to obtain an entity type set. Specifically, in this embodiment, for the propagation graph, the BERT word embedding layer can generate context-related representations, the BiLSTM layer captures bidirectional semantic dependencies, and the CRF layer optimizes entity boundaries by combining dictionary constraints (such as device number generation rules) to output an entity type set (including event, sentiment, geographic location entities, etc.) and confidence scores.

[0033] S302. Use graph attention network to learn the implicit relationships between entities in the propagation graph, and combine the business rule base to verify and complete the implicit relationships between entities to obtain the final relationships between entities; S303. Cross-validation and conflict resolution are performed on the entity attributes in the propagation graph to obtain the final entity attributes. Specifically, in this embodiment, text attributes (such as user ID) are matched with log metadata (such as device fingerprint) and image OCR results across modalities; knowledge bases such as DBpedia are queried using SPARQL (SPARQL Protocol and RDF Query Language) to verify entity attributes (such as full name / abbreviation mapping of an organization); an attribute dependency model is established based on a Bayesian network to adjudicate contradictory attributes (such as geographical location conflicts) according to data source priority (structured data > text annotation), thereby completing the attribute conflict resolution.

[0034] This embodiment significantly improves the accuracy and semantic integrity of propagation path tracing by systematically identifying, completing, and optimizing entities, relationships, and attributes in the propagation graph, providing a solid structural foundation for subsequent propagation path querying and semantic reasoning.

[0035] Furthermore, in step S3 of this embodiment, the path tracing database is stored in RDF (Resource Description Framework) format and cross-platform information dissemination tracing is performed through the SPARQL query interface.

[0036] It should be noted that storing the path tracing database in RDF format enables a structured and standardized representation of entities, events, and their semantic relationships involved in cross-platform propagation, resulting in good data scalability and interoperability. Through the SPARQL query interface, the system can perform flexible and in-depth semantic queries and path tracing based on a triple structure, supporting cross-platform and cross-modal information association retrieval and propagation path reconstruction.

[0037] After constructing the propagation map, the method further includes: S4. Based on the propagation map, construct a multi-dimensional evaluation index and normalize each evaluation index to obtain a comprehensive evaluation value of the propagation index.

[0038] Specifically, in this embodiment, the multidimensional evaluation indicators include evaluation indicators such as the breadth of dissemination (number of users covered / number of platforms), depth (dissemination level / path length), speed (dissemination peak / growth rate), sentiment tendency (sentiment distribution), and opinion (cluster analysis) to identify the mass communication effect.

[0039] In the implementation process, range standardization is used to linearly transform the multidimensional evaluation indicators, mapping the data to the [0,1] interval to eliminate dimensional differences. Simultaneously, the dispersion of the multidimensional evaluation indicators is quantified based on the entropy weight method. By calculating the probability distribution of each indicator value, information entropy is derived to reflect the amount of information. Subsequently, the objective weights of each evaluation indicator are determined based on the entropy redundancy, achieving coordinated processing of data standardization, normalization, and objective weighting to obtain normalized indicators. Finally, using the normalized indicators and specific data values, the normalized indicator z=[z1,z2,...,z] is... m The comprehensive dissemination index is obtained by weighted summation, and then mapped to the [0,100] interval through linear transformation to obtain the comprehensive evaluation value of the dissemination index. This operation ensures the comparability of the indices, and 100 points represents the theoretical maximum dissemination influence. When the dissemination influence exceeds a certain value, it indicates that a dissemination alarm event has occurred, and risk warning needs to be issued.

[0040] This embodiment enables cross-platform information dissemination analysis and possesses strong dissemination path analysis capabilities. Specifically, during implementation, this embodiment integrates heterogeneous information from cross-platform multimodal data to construct a dynamic dissemination map based on user interaction behavior and content characteristics, and achieves dissemination path tracing. This overcomes the data silo limitations of traditional methods and significantly improves the real-time performance and accuracy of cross-platform dissemination analysis, enhancing dissemination path analysis capabilities.

[0041] Example 2: This embodiment discloses a cross-platform information propagation analysis system for implementing the cross-platform information propagation analysis method in Embodiment 1; such as Figure 2 As shown, the cross-platform information dissemination analysis system includes: The data fusion module is used to collect and process multimodal data from different platforms to obtain multi-source data, and to construct a knowledge graph based on the multi-source data. The propagation graph construction module is communicatively connected to the data fusion module and is used to perform event topic mining and sentiment analysis on the knowledge graph to obtain event topic distribution information and sentiment analysis results, and construct a propagation graph based on the event topic distribution information and sentiment analysis results. The propagation tracing module is communicatively connected to the propagation graph construction module. It is used to obtain the entity type set, the relationship between entities and the entity attributes based on the propagation graph, and to construct a path tracing database based on the entity type set, the relationship between entities and the entity attributes, so as to perform cross-platform information propagation tracing based on the path tracing database.

[0042] It should be noted that the working process, working details and technical effects of the cross-platform information dissemination analysis system provided in this embodiment 2 can be found in embodiment 1, and will not be repeated here.

[0043] Example 3: Based on Embodiment 1 or 2, this embodiment discloses an electronic device, which may be a smartphone, tablet computer, laptop computer, or desktop computer, etc. The electronic device may be referred to as a user terminal, portable terminal, desktop terminal, etc. Figure 3 As shown, the electronic device includes: Memory, used to store computer program instructions; and, A processor is used to execute the computer program instructions to perform the operation of a cross-platform information propagation analysis method as described in any of Embodiment 1.

[0044] Specifically, processor 301 may include one or more processing cores, such as a quad-core processor or an octa-core processor. Processor 301 may be implemented using at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array). Processor 301 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 301 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content required to be displayed on the screen.

[0045] The memory 302 may include one or more computer-readable storage media, which may be non-transitory. The memory 302 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the non-transitory computer-readable storage media in the memory 302 are used to store at least one instruction, which is executed by the processor 301 to implement the cross-platform information propagation analysis method provided in Embodiment 1 of this application.

[0046] In some embodiments, the terminal may also optionally include a communication interface 303 and at least one peripheral device. The processor 301, memory 302, and communication interface 303 can be connected via a bus or signal line. Each peripheral device can be connected to the communication interface 303 via a bus, signal line, or circuit board. Specifically, the peripheral device includes at least one of a radio frequency circuit 304, a display screen 305, and a power supply 306.

[0047] The communication interface 303 can be used to connect at least one I / O (Input / Output) related peripheral device to the processor 301 and the memory 302. In some embodiments, the processor 301, the memory 302, and the communication interface 303 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 301, the memory 302, and the communication interface 303 can be implemented on separate chips or circuit boards, which is not limited in this embodiment.

[0048] The radio frequency (RF) circuit 304 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The RF circuit 304 communicates with communication networks and other communication devices via electromagnetic signals.

[0049] Display screen 305 is used to display the UI (User Interface). The UI may include any combination of graphics, text, icons, and video.

[0050] Power supply 306 is used to supply power to various components in electronic devices.

[0051] Example 4: Based on any one of Embodiments 1 to 3, this embodiment discloses a computer program product, including a computer program or instructions, which, when executed by a computer, implements a cross-platform information propagation analysis method as described in any one of Embodiments 1. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0052] Obviously, those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device, or fabricating them separately as individual integrated circuit modules, or fabricating multiple modules or steps as a single integrated circuit module. Thus, the present invention is not limited to any particular hardware and software combination.

[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. 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 cross-platform information dissemination analysis method, characterized in that, include: Multimodal data is collected and processed from different platforms to obtain multi-source data, and a knowledge graph is constructed based on the multi-source data. The knowledge graph is subjected to event topic mining and sentiment analysis to obtain event topic distribution information and sentiment analysis results, and a propagation graph is constructed based on the event topic distribution information and sentiment analysis results. The propagation graph is used to obtain the entity type set, the relationship between entities and the entity attributes. A path tracing database is constructed based on the entity type set, the relationship between entities and the entity attributes so as to carry out cross-platform information propagation tracing based on the path tracing database.

2. The cross-platform information dissemination analysis method according to claim 1, characterized in that, Constructing a knowledge graph based on the multi-source data includes: The multi-source data is subjected to structured parsing to obtain a structured data pool; Cross-modal feature fusion processing is performed on the media resource data in the structured data pool to obtain a joint embedding space; Entities, relationships, and attributes are extracted based on the joint embedding space, and a knowledge graph is constructed based on the entities, relationships, and attributes.

3. The cross-platform information dissemination analysis method according to claim 2, characterized in that, After forming the joint embedding space, the method further includes: A stratified sampling method is used to perform stratified sampling on the embedding representations in the joint embedding space to obtain sampled data. The quality of the sampled data is then evaluated, and if the quality evaluation fails, the embedding representations in the joint embedding space are completed to obtain a corrected joint embedding space, so as to construct a knowledge graph based on the corrected joint embedding space.

4. The cross-platform information dissemination analysis method according to claim 1, characterized in that, The knowledge graph is subjected to event topic mining and sentiment analysis to obtain event topic distribution information and sentiment analysis results, including: The BERTopic topic model is used to perform semantic clustering on the content entities in the knowledge graph to extract event topic distribution information; wherein, the event topic distribution information includes the event topic described by each content entity; The DeBERTa-v3 pre-trained model is used to perform text sentiment analysis on the content entities in the knowledge graph to obtain sentiment analysis results, which include sentiment type and intensity.

5. The cross-platform information dissemination analysis method according to claim 1, characterized in that, Based on the propagation graph, the entity type set, the relationships between entities, and the entity attributes are obtained, including: The propagation map is used to identify entity boundaries to obtain a set of entity types. The implicit relationships between entities in the propagation graph are learned using a graph attention network, and the implicit relationships between entities are verified and completed using a business rule base to obtain the final relationships between entities. Cross-validation and conflict resolution are performed on the entity attributes in the propagation graph to obtain the final entity attributes.

6. The cross-platform information dissemination analysis method according to claim 1, characterized in that, The path tracing database is stored in RDF format and uses the SPARQL query interface for cross-platform information dissemination and tracing.

7. The cross-platform information dissemination analysis method according to claim 1, characterized in that, After constructing the propagation map, the method further includes: Based on the propagation map, a multidimensional evaluation index is constructed, and each evaluation index is normalized to obtain a comprehensive evaluation value of the propagation index.

8. A cross-platform information dissemination analysis system, characterized in that, include: The data fusion module is used to collect and process multimodal data from different platforms to obtain multi-source data, and to construct a knowledge graph based on the multi-source data. The propagation graph construction module is communicatively connected to the data fusion module and is used to perform event topic mining and sentiment analysis on the knowledge graph to obtain event topic distribution information and sentiment analysis results, and construct a propagation graph based on the event topic distribution information and sentiment analysis results. The propagation tracing module is communicatively connected to the propagation graph construction module. It is used to obtain the entity type set, the relationship between entities and the entity attributes based on the propagation graph, and to construct a path tracing database based on the entity type set, the relationship between entities and the entity attributes, so as to perform cross-platform information propagation tracing based on the path tracing database.

9. An electronic device, characterized in that, include: Memory is used to store computer program instructions; as well as, A processor is configured to execute the computer program instructions to perform the operation of a cross-platform information propagation analysis method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or the instructions are executed by the computer, they implement a cross-platform information propagation analysis method as described in any one of claims 1 to 7.

Images