Modeling method of network information propagation model

Abstract

The invention relates to the technical field of information model construction, in particular to a modeling method of a network information propagation model. The method comprises the following steps: acquiring time dimension data, space dimension data and a content image; the time dimension data, the space dimension data and the content image are preprocessed, a network information propagation data set is generated, and the network information propagation data set comprises network information time dimension data, network information space dimension data and network information content dimension data; performing geographic position clustering analysis on the network information spatial dimension data to generate network information geographic position data; performing time sequence segmentation processing on the network information time dimension data to generate network information time sequence data; therefore, through refined node classification and a dynamic optimization mechanism, the problem of insufficient stability of a traditional information propagation model under various network conditions is solved, and the accuracy of the model in practical application is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of information model construction, and in particular to a modeling method of a network information propagation model. Background Art

[0002] In the current field of network information dissemination, with the continuous expansion of the scale of information diffusion and the increasing complexity of network structure, traditional information dissemination models face many challenges in dealing with massive data and dynamically changing network environments. First, the existing models rely on static indicators in evaluating node influence, which cannot fully reflect the dynamic influence of nodes in different time and space dimensions, which leads to insufficient accuracy of the model in predicting information dissemination paths and rates. Secondly, in the simulation process of information diffusion, traditional methods usually adopt a single propagation mechanism and lack support for diversified propagation modes, which makes the model show poor adaptability when facing multiple information diffusion scenarios. In addition, the existing models rely on preset rules in the selection of information diffusion paths and lack effective modeling of dynamic interaction relationships between nodes, making it difficult to accurately simulate information dissemination behavior in complex networks. Furthermore, the setting of model parameters usually relies on manual experience and lacks an adaptive adjustment mechanism, resulting in limited generalization ability in different application scenarios. Therefore, there is an urgent need for a network information dissemination model that can dynamically evaluate node influence, support multiple diffusion mechanisms, and have adaptive parameter adjustment capabilities to improve the accuracy and adaptability of the model in practical applications. Summary of the invention

[0003] Based on this, it is necessary to provide a modeling method for a network information propagation model to solve at least one of the above technical problems.

[0004] To achieve the above object, a modeling method for a network information propagation model is provided, the method comprising the following steps:

[0005] Step S1: obtaining time dimension data, space dimension data and content image; preprocessing the time dimension data, space dimension data and content image to generate a network information propagation data set, wherein the network information propagation data set includes network information time dimension data, network information space dimension data and network information content dimension data;

[0006] Step S2: Performing geographic location clustering analysis on network information space dimension data to generate network information geographic location data; performing time series segmentation processing on network information time dimension data to generate network information time series data; performing semantic content feature extraction on network information content dimension data to generate network information semantic data; performing propagation structure construction on network information geographic location data, network information time series data and network information semantic data to generate network information propagation structure data;

[0007] Step S3: Based on the preset network information influence model, the network information geographic location data, network information time series data and network information semantic data are evaluated for influence factors to generate network information node influence data; the network information node influence data is simulated for propagation probability using the dynamic propagation probability function to generate a dynamic propagation path; the network information node influence data and the dynamic propagation path are mapped to the network information model to generate a preliminary information propagation model;

[0008] Step S4: Simulate the information diffusion pattern of the preliminary information propagation model to generate network information simulation data; adjust the model parameters of the preliminary information propagation model based on the network information simulation data to generate a network information propagation model.

[0009] The beneficial effect of the present invention is that by acquiring and preprocessing time dimension data, space dimension data and content images, a network information propagation data set with spatiotemporal dynamics and content characteristics is generated. This process effectively integrates data of different dimensions and lays the foundation for the construction of subsequent information propagation models. Through geographic location clustering analysis, time series segmentation processing and semantic content feature extraction, the complex information in the original data is hierarchically structured, and geographic location data, time series data and semantic data are generated respectively, which not only ensures the consistency and integrity of the data in space, time and content, but also improves the availability and processing efficiency of the data. Using the propagation structure construction step, the above-mentioned dimensional data are fused to generate network information propagation structure data. This step reveals the inherent laws and structural characteristics of information propagation through cross-analysis of geographic, time and semantic dimensions, and lays a solid foundation for subsequent impact factor evaluation. In the impact factor evaluation link, the multidimensional data is quantitatively analyzed through the preset network information influence model to generate network information node influence data, thereby revealing the key role of each node in the propagation process. The propagation probability simulation of node influence data using dynamic propagation probability function not only reflects the dynamic characteristics of information propagation, but also provides the ability to predict the actual propagation path through the construction of a preliminary information propagation model. Finally, the final network information propagation model is generated through simulation and parameter adjustment of the preliminary information propagation model. The core of this step is to adjust the feedback based on the simulation data, so that the model has higher accuracy and adaptability, and can better reflect the complex changes in the actual propagation process. Therefore, the present invention solves the problem of insufficient stability of traditional information propagation models under various network conditions through refined node classification and dynamic optimization mechanism, and improves the accuracy of the model in practical applications.

[0010] Preferably, step S1 comprises the following steps:

[0011] Step S11: Acquire time dimension data, space dimension data and content picture data;

[0012] Step S12: extracting content data from the content image using OCR technology to generate content dimension data;

[0013] Step S13: performing time data discretization processing on the time dimension data to generate network information time dimension data; performing coordinate system one processing on the space dimension data to generate network information space dimension data;

[0014] Step S14: performing text cleaning on the content dimension data to generate content text cleaning data; performing keyword extraction on the content text cleaning data to generate text content keyword data; performing text vectorization processing on the text content keyword data to generate network information content dimension data;

[0015] Step S15: Merge the network information time dimension data, the network information space dimension data and the network information content dimension data to generate a network information propagation data set.

[0016] The present invention ensures the diversity and integrity of data sources by acquiring time dimension data, space dimension data and content image data, and provides multi-dimensional input for subsequent processing. The content image is extracted by using OCR technology to effectively convert the unstructured information in the image into analyzable text data, thereby generating structured content dimension data and improving the depth and breadth of data processing. The time dimension data is processed by time data discretization to generate network information time dimension data suitable for analysis. This process discretizes continuous time information, which is convenient for subsequent time series analysis. The space dimension data is processed by coordinate system one, eliminating the differences between different coordinate systems, generating standardized network information space dimension data, and ensuring the accuracy and consistency of spatial data. The content dimension data is cleaned and keyword extracted, and the generated text content keyword data is converted into a high-dimensional numerical data form by vectorization processing, generating refined network information content dimension data. The application of the text cleaning step effectively removes noise and redundant information, improves the purity of the data and the accuracy of the analysis; the keyword extraction captures the core information of the text, and provides efficient data support for subsequent analysis. Finally, the network information time dimension data, network information space dimension data and network information content dimension data were merged to generate a complete network information dissemination data set. This data set combines information in the three dimensions of time, space and content to form a multi-level structured data set, providing a comprehensive data foundation for the study of network information dissemination laws. Such integrated processing also improves the consistency and availability of data, ensuring the accuracy and reliability of subsequent analysis.

[0017] Preferably, step S2 comprises the following steps:

[0018] Step S21: Identify information clusters on the network information spatial dimension data to generate network information cluster data; locate hotspots on the network information cluster data to generate network information hotspot location data; cluster geographical locations on the network information hotspot location data to generate network information geographical location data;

[0019] Step S22: performing periodic time series recognition on the network information time dimension data to generate network information periodic data; performing time series segmentation processing on the network information periodic data to generate network information time series data;

[0020] Step S23: Calculate the word weights of the network information content dimension data to generate network information word weight data; perform semantic feature extraction processing on the network information word weight data to generate network information semantic feature data; perform semantic data generation on the network information semantic feature data to generate network information semantic data;

[0021] Step S24: constructing a propagation structure for the network information geographic location data, the network information time series data and the network information semantic data to generate network information propagation structure data.

[0022] The present invention generates network information geographic location data through a series of processing steps including information gathering area identification, hot spot location and geographic location clustering. This process not only successfully identifies the key geographic areas in information dissemination, but also reveals the geographical relationship between these hot spots through cluster analysis, providing a detailed data basis for understanding the spatial diffusion path of information. In the time dimension, network information time series data is generated through periodic time series identification and time series segmentation processing. This processing step captures the periodic laws and time characteristics in information dissemination, converts the original time data into time series data with more analytical value, enhances the time series analysis ability of the data, and provides time dimension support for the subsequent construction of the dissemination structure. Content dimension data generates network information semantic data through the steps of term weight calculation, semantic feature extraction and semantic data generation. Term weight calculation effectively measures the importance of each term in the text, and semantic feature extraction further mines the core semantic information of the text, and through semantic data generation, structures the text data into numerical data with semantic depth, thereby improving the accuracy and richness of content analysis. Finally, network information dissemination structure data is generated through the dissemination structure construction of network information geographic location data, time series data and semantic data. This dataset not only integrates information in the three dimensions of space, time and content, but also reveals the interactive influence and propagation paths of information in different dimensions through the construction of propagation structure, providing the foundation for a multi-dimensional and multi-level network information propagation model.

[0023] Preferably, step S24 includes the following steps:

[0024] Step S241: aligning the network information geographic location data, the network information time series data and the network information semantic data to generate a network information alignment data set;

[0025] Step S242: time-position tagging is performed on the network information geographic location data and the network information time series data to generate network time-position path change data; a propagation node graph is designed for the network time-position path change data based on the graph structure, and node attributes are attached using the network information semantic data to generate a network information node structure graph;

[0026] Step S243: Perform community detection processing on the network information node structure diagram and the network information alignment data set to generate network information dissemination structure data.

[0027] The present invention generates a unified network information alignment data set by aligning network information geographic location data, time series data and semantic data. This process ensures the consistency of data of different dimensions in time and space, provides a comprehensive data basis for subsequent analysis, eliminates the differences in format, scale and time tagging of multi-dimensional data, and improves the accuracy and reliability of data processing. The network information geographic location data and time series data are marked with time-position to generate network time-position path change data. Through this marking step, the change path of information in time and space is systematically recorded, providing detailed data support for understanding the propagation dynamics of information in different geographical regions and time periods. The propagation node graph designed based on the graph structure not only visualizes the time-position path change data as a graph structure, but also gives each node rich semantic attributes by adding network information semantic data. This process greatly improves the meticulousness of propagation path analysis, making the relationship between nodes and information flow in the graph structure clearer. Finally, the network information propagation structure data is generated by performing community detection processing on the network information node structure graph and the network information alignment data set. This step identifies highly correlated node groups in the information dissemination process through community detection, revealing the community structure characteristics of information dissemination. Community processing not only helps identify key dissemination nodes, but also reveals the aggregation and diffusion patterns of information in the network, providing profound insights into understanding the inherent laws of information dissemination.

[0028] Preferably, step S3 comprises the following steps:

[0029] Step S31: Evaluate the influence factors of network information geographic location data, network information time series data and network information semantic data based on a preset network information influence model to generate network information node influence data;

[0030] Step S32: using a dynamic propagation probability function to simulate the propagation probability of the network information node influence data to generate a dynamic propagation path;

[0031] Step S33: constructing an information diffusion mechanism based on the network information node influence data and the dynamic propagation path to generate a network information diffusion mechanism;

[0032] Step S34: Map the network information node influence data, dynamic propagation path and network information diffusion mechanism to the network information model to generate a preliminary information propagation model.

[0033] The present invention evaluates the influence factors of geographic location, time series and semantic data based on a preset network information influence model, and generates network information node influence data. This process converts complex multidimensional data into easy-to-understand node influence metrics through quantitative analysis, providing key data support for identifying key nodes in the network. Through influence evaluation, it is further clarified which nodes play a vital role in the information propagation process, thereby laying a solid foundation for subsequent propagation simulation. The network information node influence data is simulated by propagation probability using a dynamic propagation probability function to generate a dynamic propagation path. This simulation step accurately predicts the propagation path of information in the network through the application of probability functions, and dynamically displays the flow process of information between different nodes. In this way, not only can the spatiotemporal dynamic characteristics of information propagation be captured, but also the actual contribution of node influence to information diffusion can be reflected. This step significantly improves the adaptability of the model to complex networks by simulating the information diffusion behavior in real scenarios at the data level. The information diffusion mechanism is constructed based on the node influence data and the dynamic propagation path to generate a network information diffusion mechanism. Through the construction of this mechanism, the propagation logic and rules of information between different nodes are clarified, thereby revealing the inherent mechanism of information diffusion. The establishment of the diffusion mechanism not only reflects the macro pattern of information flow, but also provides microscopic insights into the information diffusion process, and can classify and predict different types of network information diffusion. The construction of this mechanism makes the model more explanatory and predictive, providing strong support for the study of complex information networks. The node influence data, dynamic propagation paths, and diffusion mechanisms are mapped to the network information model to generate a preliminary information propagation model. This mapping process successfully constructs a multi-level information propagation model covering the spatiotemporal and semantic dimensions by integrating the results of the previous steps into a comprehensive model. This preliminary model not only has high accuracy, but can also dynamically adapt to changes in different propagation environments, thus providing a powerful tool for the comprehensive understanding and control of information propagation.

[0034] Preferably, step S31 includes the following steps:

[0035] Step S311: evaluating the influence factors of the network information geographic location data, network information time series data and network information semantic data based on a preset network information influence model, and generating a geographic location factor, a time series factor and a semantic information factor;

[0036] Step S312: Mapping the geographical location factor, the time series factor and the semantic information factor for information propagation association to generate a network information factor association analysis result; assigning weighted parameters to the network information factor association analysis result to generate network information factor parameter weighted data;

[0037] Step S313: Use the fuzzy comprehensive evaluation method to divide the membership of the network information factor association analysis results to generate network information membership data; quantify the influence of the network information factor parameter weighted data and the network information membership data to generate network information node influence data.

[0038] The present invention uses a preset network information influence model to evaluate the influence factors of geographic location, time series and semantic information, and generates geographic location factors, time series factors and semantic information factors respectively. This evaluation process is based on data, decomposes the influence of various types of information into specific factors, and can capture the influence of different dimensional information on the overall network information dissemination at a subtle level, thereby providing accurate basic data for subsequent analysis. These factors are mapped to information dissemination associations to generate network information factor association analysis results. Through this association mapping step, factors of different dimensions are integrated into an association analysis framework, revealing the complex relationship and interaction between geographic location, time series and semantic information. This integration not only improves the comprehensiveness of information dissemination analysis, but also provides a deep insight into the causal chain of network information dissemination. Weighted parameters are assigned to the association analysis results to generate network information factor parameter weighted data. This process further refines the relative influence between factors by weighting the importance of each factor, ensuring the accuracy and reliability of subsequent analysis. The network information factor association analysis results are divided into membership by fuzzy comprehensive evaluation method to generate network information membership data. The application of fuzzy comprehensive evaluation method ensures the rationality and scientificity of membership division by processing complex multidimensional data with fuzzy logic. The generation of membership data marks the quantitative processing of fuzzy relationships between information factors, which can not only effectively identify key nodes in information dissemination, but also provide a scientific basis for further quantitative evaluation. Finally, the network information node influence data is generated by quantifying the influence of network information factor parameter weighted data and membership data. This influence quantification step combines the weighted and membership information of different factors to accurately quantify the influence of each node in network information dissemination. The generated node influence data provides key data support for the construction and optimization of network information dissemination models, which can greatly improve the prediction accuracy of information dissemination paths and the adaptability of models.

[0039] Preferably, step S32 includes the following steps:

[0040] Step S321: Performing propagation path potential analysis on the network information node influence data to generate bright node potential analysis data; performing potential propagation link identification on the node potential analysis data to generate potential node analysis data;

[0041] Step S322: Perform dynamic spatiotemporal mapping comparison on the bright node potential analysis data and the latent node analysis data to generate dynamic propagation mapping data;

[0042] Step S323: Use the dynamic propagation probability function to perform propagation probability simulation on the dynamic propagation mapping data to generate a dynamic propagation path.

[0043] The present invention generates the potential analysis data of the node by performing the propagation path potential analysis on the network information node influence data. This analysis step deeply explores the potential of the node in the information propagation path and clarifies which nodes will become the key propagation nodes in the future propagation process. This potential analysis provides a substantial basis for identifying the potential key path of information propagation at the data level, ensuring the accuracy and operability of the subsequent propagation path simulation. Potential propagation links are identified for the node potential analysis data to generate potential node analysis data. By identifying potential propagation links, this step further deepens the understanding of the information propagation path, especially in complex network structures, which can not only discover the propagation path between explicit nodes, but also reveal the potential propagation links hidden under the data surface. The generated potential node analysis data provides a multi-dimensional perspective for the accurate simulation of the dynamic propagation path, making the prediction of the information propagation path more comprehensive and profound. The potential analysis data of the node and the potential node analysis data are compared by dynamic spatiotemporal mapping to generate dynamic propagation mapping data. Through this dynamic spatiotemporal mapping comparison process, the dynamic change factors of time and space are combined to comprehensively analyze the propagation potential of different nodes in different spatiotemporal dimensions. This mapping data not only shows the dynamic characteristics of information propagation in the space-time dimension, but also provides data support for understanding the space-time evolution law in the process of information propagation, and improves the space-time accuracy of propagation path prediction. Finally, the dynamic propagation probability function is used to simulate the propagation probability of the dynamic propagation mapping data to generate a dynamic propagation path. This step converts complex dynamic mapping data into specific propagation path predictions through probability simulation, which provides a clearer description of the flow process of information in the network. The generation of dynamic propagation paths provides a highly targeted and practical propagation path prediction tool at the data level, so that in practical applications, the dynamic changes of information propagation can be better grasped, thereby optimizing information propagation strategies.

[0044] Preferably, step S33 includes the following steps:

[0045] Step S331: classifying the network information node influence data by node role to generate a network information node classification result, wherein the network information node classification result includes core nodes, edge nodes and relay nodes;

[0046] Step S332: construct a robust core for the core node to generate a core preferred link; connect redundant nodes in series on the edge node to generate an edge redundant link; optimize the transmission link on the relay node to generate a relay node link;

[0047] Step S333: Use network flow technology to construct an information diffusion mechanism for core preferred links, edge redundant links, relay node links and dynamic propagation paths to generate a network information diffusion mechanism;

[0048] The present invention generates network information node classification results by classifying the node roles of network information node influence data, and these results include core nodes, edge nodes and relay nodes. This classification process systematically divides the nodes in the network according to their roles in the information dissemination process, which helps to clarify the functional roles of each node in information diffusion. This classification lays the foundation for subsequent optimization processing, so that more refined optimization strategies can be adopted for nodes with different roles, thereby improving the propagation efficiency of the overall network. The core nodes are robustly constructed to generate core preferred links. Through robustness analysis, the links of the core nodes are optimized to improve their stability and anti-interference ability in information dissemination. This optimization ensures that the core nodes can still effectively maintain the continuity and efficiency of information dissemination when facing faults and attacks in the network. In addition, redundant nodes are connected in series on the edge nodes to generate edge redundant links. This step improves the redundancy and reliability of information dissemination by establishing redundant links between edge nodes, and can effectively avoid information dissemination interruptions caused by single point failures, thereby improving the overall stability of the network. For relay nodes, transmission link optimization is performed to generate relay node links. The optimization of relay nodes improves the efficiency and quality of data transmission and ensures the smooth flow of information in the network. Finally, network flow technology is used to construct an information diffusion mechanism for core preferred links, edge redundant links, relay node links and dynamic propagation paths, generating a network information diffusion mechanism. The construction of this mechanism forms a comprehensive diffusion mechanism model through a comprehensive analysis of various links and propagation paths in the network, which can accurately control the information diffusion process at the data level. Through the application of network flow technology, the efficient propagation of information in the network and the optimal allocation of resources are ensured, making the information diffusion process not only more efficient, but also able to maintain stability when the network load changes and the node status changes.

[0049] Preferably, step S34 includes the following steps:

[0050] Step S341: Establish an inter-node strength matrix for network information node influence data to generate a network information inter-node strength matrix; construct a propagation link weight matrix for the dynamic propagation path to generate a network information propagation link weight matrix; use the information diffusion mechanism as a dynamic weight parameter matrix to generate a network information dynamic weight parameter matrix;

[0051] Step S342: performing multiple matrix products on the network information node strength matrix, the network information transmission link weight matrix, and the network information dynamic weight parameter matrix to generate a network information transmission data matrix;

[0052] Step S343: Use the node influence data as the initial model parameters to generate a model for the network information propagation data matrix to generate a preliminary information propagation model.

[0053] The present invention generates a network information node intensity matrix by establishing an inter-node intensity matrix for network information node influence data. This matrix reflects the interaction intensity and influence relationship between nodes in the network, and provides basic data for subsequent propagation link weight analysis. The propagation link weight matrix is ​​constructed for the dynamic propagation path, and a network information propagation link weight matrix is ​​generated, which describes the weight distribution of information in the propagation link, and further clarifies the priority and resource allocation in the information transmission process. The information diffusion mechanism is generated as a dynamic weight parameter matrix to form a network information dynamic weight parameter matrix. This matrix takes into account the dynamic characteristics of information diffusion and provides a basis for real-time adjustment of the flow and distribution of information in the network. Multiple matrix product operations comprehensively consider the network information node intensity matrix, the network information propagation link weight matrix and the network information dynamic weight parameter matrix to generate a network information propagation data matrix. This process obtains a more comprehensive and accurate network information propagation data matrix by integrating different types of data matrices. The matrix comprehensively considers the interaction intensity between nodes, the weight of the propagation link and the dynamic weight adjustment, so that the data performance of network information propagation is richer and more accurate. Finally, the node influence data is used as the initial model parameter to generate the model for the generated network information propagation data matrix, thus generating a preliminary information propagation model. This model uses node influence data and propagation data matrix in a data-driven way to form an information propagation model that can be used in practical applications. Through this model, the information propagation process in the network is simulated and predicted, providing an effective tool for network optimization and strategy adjustment.

[0054] Preferably, step S4 comprises the following steps:

[0055] Step S41: performing multi-information diffusion mode simulation on the preliminary information propagation model to generate network information simulation data, wherein the network information simulation data includes network information pressure simulation data and network information normal flow data;

[0056] Step S42: tracing the flow information pressure source of the network information pressure simulation data to generate pressure simulation test cause data; adjusting the influence factor weight of the network information node influence data based on the pressure simulation test cause data to generate network information influence adjustment data; performing network flow information detection on the network information normal flow data based on a preset threshold to generate network information propagation detection data;

[0057] Step S43: adjusting model parameters of the preliminary information propagation model based on the network information influence adjustment data and the network information propagation detection data, thereby generating a network information propagation model.

[0058] The present invention generates network information simulation data by simulating multiple information diffusion modes on the preliminary information propagation model, including network information pressure simulation data and network information normal flow data. This simulation process provides performance data of network information under various conditions by simulating different propagation modes. These data include network performance under pressure conditions (network information pressure simulation data) and network performance under normal flow conditions (network information normal flow data). Through such simulation, the performance of the information propagation model under normal and extreme conditions can be fully understood, thereby providing basic data for subsequent model optimization. The network information pressure simulation data is traced for flow information pressure, and pressure simulation test cause data is generated. This step determines the specific cause of the pressure through tracing analysis, which helps to understand the bottleneck in the network under pressure conditions. Based on these test cause data, the influence factor weight of the network information node influence data is adjusted to generate network information influence adjustment data. This adjustment process ensures that the performance of the network node under different pressure conditions can better reflect the actual situation by optimizing the node influence, thereby improving the robustness and adaptability of the model. Based on the preset threshold, the network information normal flow data is tested for network flow information, and network information propagation detection data is generated. This detection process ensures the stability of network traffic under normal conditions and identifies existing traffic anomalies and problems. Based on the network information influence adjustment data and network information propagation detection data, the model parameters of the preliminary information propagation model are adjusted to generate a network information propagation model. Through this adjustment, the model can more accurately reflect the information propagation situation in the actual network and optimize the performance of the model under different conditions. This process realizes the dynamic adjustment and optimization of the model through comprehensive analysis of simulation data and detection data, so that the final generated network information propagation model can better adapt to the actual network environment and improve the practical application value of the model.

[0059] The beneficial effect of the present invention is that by acquiring and preprocessing time dimension data, space dimension data and content images, a network information propagation data set with spatiotemporal dynamics and content characteristics is generated. This process effectively integrates data of different dimensions and lays the foundation for the construction of subsequent information propagation models. Through geographic location clustering analysis, time series segmentation processing and semantic content feature extraction, the complex information in the original data is hierarchically structured, and geographic location data, time series data and semantic data are generated respectively, which not only ensures the consistency and integrity of the data in space, time and content, but also improves the availability and processing efficiency of the data. Using the propagation structure construction step, the above-mentioned dimensional data are fused to generate network information propagation structure data. This step reveals the inherent laws and structural characteristics of information propagation through cross-analysis of geographic, time and semantic dimensions, and lays a solid foundation for subsequent impact factor evaluation. In the impact factor evaluation link, the multidimensional data is quantitatively analyzed through the preset network information influence model to generate network information node influence data, thereby revealing the key role of each node in the propagation process. The propagation probability simulation of node influence data using dynamic propagation probability function not only reflects the dynamic characteristics of information propagation, but also provides the ability to predict the actual propagation path through the construction of a preliminary information propagation model. Finally, the final network information propagation model is generated through simulation and parameter adjustment of the preliminary information propagation model. The core of this step is to adjust the feedback based on the simulation data, so that the model has higher accuracy and adaptability, and can better reflect the complex changes in the actual propagation process. Therefore, the present invention solves the problem of insufficient stability of traditional information propagation models under various network conditions through refined node classification and dynamic optimization mechanism, and improves the accuracy of the model in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic diagram of the steps of a modeling method for a network information propagation model;

[0061] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0062] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0063] Figure 4 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0064] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0065] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0066] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0067] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0068] To achieve this, please refer to Figures 1 to 4 , a modeling method for a network information propagation model, the method comprising the following steps:

[0069] Step S1: obtaining time dimension data, space dimension data and content image; preprocessing the time dimension data, space dimension data and content image to generate a network information propagation data set, wherein the network information propagation data set includes network information time dimension data, network information space dimension data and network information content dimension data;

[0070] Step S2: Performing geographic location clustering analysis on network information space dimension data to generate network information geographic location data; performing time series segmentation processing on network information time dimension data to generate network information time series data; performing semantic content feature extraction on network information content dimension data to generate network information semantic data; performing propagation structure construction on network information geographic location data, network information time series data and network information semantic data to generate network information propagation structure data;

[0071] Step S3: Based on the preset network information influence model, the network information geographic location data, network information time series data and network information semantic data are evaluated for influence factors to generate network information node influence data; the network information node influence data is simulated for propagation probability using the dynamic propagation probability function to generate a dynamic propagation path; the network information node influence data and the dynamic propagation path are mapped to the network information model to generate a preliminary information propagation model;

[0072] Step S4: Simulate the information diffusion pattern of the preliminary information propagation model to generate network information simulation data; adjust the model parameters of the preliminary information propagation model based on the network information simulation data to generate a network information propagation model.

[0073] The beneficial effect of the present invention is that by acquiring and preprocessing time dimension data, space dimension data and content images, a network information propagation data set with spatiotemporal dynamics and content characteristics is generated. This process effectively integrates data of different dimensions and lays the foundation for the construction of subsequent information propagation models. Through geographic location clustering analysis, time series segmentation processing and semantic content feature extraction, the complex information in the original data is hierarchically structured, and geographic location data, time series data and semantic data are generated respectively, which not only ensures the consistency and integrity of the data in space, time and content, but also improves the availability and processing efficiency of the data. Using the propagation structure construction step, the above-mentioned dimensional data are fused to generate network information propagation structure data. This step reveals the inherent laws and structural characteristics of information propagation through cross-analysis of geographic, time and semantic dimensions, and lays a solid foundation for subsequent impact factor evaluation. In the impact factor evaluation link, the multidimensional data is quantitatively analyzed through the preset network information influence model to generate network information node influence data, thereby revealing the key role of each node in the propagation process. The propagation probability simulation of node influence data using dynamic propagation probability function not only reflects the dynamic characteristics of information propagation, but also provides the ability to predict the actual propagation path through the construction of a preliminary information propagation model. Finally, the final network information propagation model is generated through simulation and parameter adjustment of the preliminary information propagation model. The core of this step is to adjust the feedback based on the simulation data, so that the model has higher accuracy and adaptability, and can better reflect the complex changes in the actual propagation process. Therefore, the present invention solves the problem of insufficient stability of traditional information propagation models under various network conditions through refined node classification and dynamic optimization mechanism, and improves the accuracy of the model in practical applications.

[0074] In the embodiment of the present invention, reference Figure 1The above is a schematic diagram of the steps of a modeling method of a network information propagation model of the present invention. In this example, the modeling method of a network information propagation model includes the following steps:

[0075] Step S1: obtaining time dimension data, space dimension data and content image; preprocessing the time dimension data, space dimension data and content image to generate a network information propagation data set, wherein the network information propagation data set includes network information time dimension data, network information space dimension data and network information content dimension data;

[0076] In the embodiment of the present invention, the process of obtaining time dimension data, space dimension data and content image uses a variety of data acquisition technologies, such as timestamp recording, geographic location tagging and image capture. The time dimension data comes from the time record of the event, the space dimension data is obtained through the geographic information system (GIS) and positioning technology, and the content image is captured by the image acquisition device. These raw data are preprocessed to improve the quality and consistency of the data. The time dimension data needs to be standardized in time format and missing values ​​are filled to ensure that it can accurately reflect the time series characteristics in the analysis. The space dimension data needs to be converted by the coordinate system and smoothed by the geographic data to eliminate errors and unify the data format. The content image is cleaned by image processing technology, including noise removal, image scaling and format conversion, for subsequent analysis. Through these preprocessing steps, a network information dissemination data set is generated. The data set includes network information time dimension data, network information space dimension data and network information content dimension data, where each data type is refined. The time dimension data of network information records the time series information of events, providing a basis for analyzing the temporal changes of network information; the spatial dimension data of network information provides the geographical distribution of information, providing a basis for spatial analysis and geographical association; the content dimension data of network information includes key information extracted from content images, such as text information and image features, which is used for content analysis and information extraction.

[0077] Step S2: Performing geographic location clustering analysis on network information space dimension data to generate network information geographic location data; performing time series segmentation processing on network information time dimension data to generate network information time series data; performing semantic content feature extraction on network information content dimension data to generate network information semantic data; performing propagation structure construction on network information geographic location data, network information time series data and network information semantic data to generate network information propagation structure data;

[0078] In the embodiment of the present invention, geographic location clustering analysis is performed on the spatial dimension data of network information, and clustering algorithms in spatial data analysis, such as K-means clustering, are applied. These algorithms group the spatial dimension data by geographic location, identify geographically close areas, and thus generate network information geographic location data. This process can reveal the data concentration trend in the geographic location, and provide a basis for further geographic information analysis. Time series segmentation processing is performed on the time dimension data of network information, and time series analysis techniques, such as the sliding window method, are used. These techniques divide the time series data into multiple time periods to facilitate the identification and analysis of temporal change patterns, thereby generating network information time series data. The processing of network information content dimension data includes semantic content feature extraction, and this process uses natural language processing (NLP) technology, such as word embedding, to extract key semantic features from content images. These technologies can convert image content into analyzable semantic data, generate network information semantic data, and provide deep feature information for content-related analysis. Finally, the propagation structure of network information geographic location data, network information time series data, and network information semantic data is constructed, and graph theory and network analysis methods are applied. These methods combine geographic location, time series and semantic data by building a graph structure of nodes and edges to form a network information propagation structure data. This step builds a comprehensive propagation structure model by combining data of different dimensions.

[0079] Step S3: Based on the preset network information influence model, the network information geographic location data, network information time series data and network information semantic data are evaluated for influence factors to generate network information node influence data; the network information node influence data is simulated for propagation probability using the dynamic propagation probability function to generate a dynamic propagation path; the network information node influence data and the dynamic propagation path are mapped to the network information model to generate a preliminary information propagation model;

[0080] In the embodiment of the present invention, based on the preset network information influence model, the influence factor evaluation is performed on the network information geographic location data, network information time series data and network information semantic data. This process uses a multi-dimensional influence analysis method. This includes calculating the influence of each node in the information propagation process through a weighted algorithm. These methods take into account the concentration of geographic location, the change trend in the time series and the relevance of semantic content, and generate network information node influence data. This data reveals the relative importance and role of each node in the entire network, and provides a basis for subsequent propagation analysis. The network information node influence data is simulated for propagation probability using a dynamic propagation probability function. This process uses dynamic models in probability theory and statistics, such as Markov processes, to simulate the propagation path of information in the network by dynamically adjusting the propagation probability function. These dynamic propagation models take into account the changes in node influence and its impact on information diffusion, thereby generating a dynamic propagation path. This step can predict the actual propagation route of information in the network and estimate its propagation effect. The network information node influence data and the dynamic propagation path are mapped to the network information model, and a model mapping technology is used, as shown in the model mapping. These techniques generate preliminary information diffusion models by mapping influence data and diffusion paths into a comprehensive model.

[0081] Step S4: Simulate the information diffusion pattern of the preliminary information propagation model to generate network information simulation data; adjust the model parameters of the preliminary information propagation model based on the network information simulation data to generate a network information propagation model.

[0082] In an embodiment of the present invention, the information diffusion pattern of the preliminary information propagation model is simulated, and this process uses computer simulation technology and dynamic system analysis methods. Specifically, simulation algorithms such as Monte Carlo simulation are used. These algorithms can simulate the diffusion process of information under different conditions and generate network information simulation data. These simulation technologies take into account the influence of each node in the network and the dynamic changes of the propagation path. Through repeated simulation and statistical analysis, detailed network information simulation data are generated. The main goal of this step is to verify the performance of the preliminary information propagation model under various assumptions to ensure its reliability in practical applications. Based on the generated network information simulation data, the model parameters of the preliminary information propagation model are adjusted. This process parameter optimization technology, such as optimization algorithm, adjusts the parameter settings in the model by comparing the simulation data with the actual observation data to reduce errors and improve prediction accuracy. Specific technologies include parameter estimation, sensitivity analysis and error correction. These technologies can refine and adjust the key parameters in the model to ensure that the model more accurately reflects the actual situation of information propagation. By adjusting the model parameters, an optimized network information propagation model is generated.

[0083] Preferably, step S1 comprises the following steps:

[0084] Step S11: Acquire time dimension data, space dimension data and content picture data;

[0085] Step S12: extracting content data from the content image using OCR technology to generate content dimension data;

[0086] Step S13: performing time data discretization processing on the time dimension data to generate network information time dimension data; performing coordinate system one processing on the space dimension data to generate network information space dimension data;

[0087] Step S14: performing text cleaning on the content dimension data to generate content text cleaning data; performing keyword extraction on the content text cleaning data to generate text content keyword data; performing text vectorization processing on the text content keyword data to generate network information content dimension data;

[0088] Step S15: Merge the network information time dimension data, the network information space dimension data and the network information content dimension data to generate a network information propagation data set.

[0089] In the embodiment of the present invention, time dimension data, space dimension data and content image data are obtained. This process uses sensors and database queries to collect raw data from multiple sources. These data provide basic raw materials for subsequent processing and cover all aspects required for information dissemination analysis. The content image is extracted using OCR (Optical Character Recognition) technology to generate content dimension data. OCR technology converts the text in the image into processable character data through an image recognition algorithm. The technologies used include convolutional neural network (CNN) and deep learning models to ensure high-accuracy extraction of text. The content image is converted into structured text data, which lays the foundation for further text processing and analysis. The time dimension data is discretized to generate network information time dimension data. This process segments and discretizes the time series data to make the data more suitable for the analysis model. The method used includes time window segmentation and discretization algorithm to convert continuous time data into discrete time period data. At the same time, the space dimension data is processed in a unified coordinate system. Through coordinate standardization, network information space dimension data is generated to ensure the comparability and consistency of spatial data under a unified coordinate system. The content dimension data is cleaned up to generate content text cleaned up data. This process includes removing noise, punctuation and irrelevant information, and the cleaned data is convenient for subsequent analysis. Through keyword extraction technology, such as TF-IDF (Term Frequency-Inverse Document Frequency), text content keyword data is generated, and further text vectorization processing is performed, such as using word embedding technology (Word Embeddings), the text is converted into a vector representation, and network information content dimension data is generated, so that the text data can be converted into a form suitable for calculation and modeling. Finally, the network information time dimension data, network information space dimension data and network information content dimension data are merged. This process uses data fusion technology, such as data splicing, to integrate data of each dimension and generate a network information dissemination data set. After the data is merged, a comprehensive database is formed, which contains information on time, space and content.

[0090] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0091] Step S21: Identify information clusters on the network information spatial dimension data to generate network information cluster data; locate hotspots on the network information cluster data to generate network information hotspot location data; cluster geographical locations on the network information hotspot location data to generate network information geographical location data;

[0092] Step S22: performing periodic time series recognition on the network information time dimension data to generate network information periodic data; performing time series segmentation processing on the network information periodic data to generate network information time series data;

[0093] Step S23: Calculate the word weights of the network information content dimension data to generate network information word weight data; perform semantic feature extraction processing on the network information word weight data to generate network information semantic feature data; perform semantic data generation on the network information semantic feature data to generate network information semantic data;

[0094] Step S24: constructing a propagation structure for the network information geographic location data, the network information time series data and the network information semantic data to generate network information propagation structure data.

[0095] In an embodiment of the present invention, the information clustering area identification technology identifies the information clustering area through density clustering of spatial data and generates network information clustering area data. This process uses spatial clustering algorithms such as DBSCAN (Density-Based Spatial Clustering of Applications with Noise) to determine high-density areas, thereby identifying clustering patterns in the data. By locating the hot spots of the information clustering area data, the hot spots in the data are located using hot spot analysis technology (such as heat map analysis). This step further refines the clustering area into a hot spot area. The hot spot location data is further clustered by geographic location, and the geographic location clustering technology (distance-based clustering) is applied to generate network information geographic location data. These technologies help identify and map the characteristics of information dissemination and its distribution in the geographic location. Periodic time series recognition is performed on the network information time dimension data, and time series analysis technology (such as Fourier transform, periodic analysis) is used to extract periodic patterns in the data to generate network information periodic data. The network information cycle data is processed by time series segmentation, and the time series decomposition technology (such as STL decomposition) is applied to divide the time series data into trend, seasonal and residual parts, thereby generating network information time series data. These technologies support in-depth understanding of the structure of time series data and identification of temporal change patterns. The term weight calculation uses the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm to process the content dimension data and generate network information term weight data. This process evaluates the importance of terms in the text and calculates the relative importance of each term through the weight assignment method. The semantic feature extraction of the term weight data is processed, and the word embedding technology (such as Word2Vec) is used to capture the semantic information of the text to generate network information semantic feature data. This process converts the contextual information of the term into a higher-dimensional semantic feature representation. The network information semantic feature data is further processed to generate network information semantic data, aiming to extract meaningful semantic content from the text data and support more in-depth analysis. Finally, the propagation structure of network information geographic location data, network information time series data and network information semantic data is constructed. This process integrates geographic, temporal and semantic data, and describes the propagation path and structure of information by constructing a propagation network model (a network model based on graph theory). These technical means combine the geographical distribution, temporal changes and semantic characteristics of data to form network information propagation structure data.

[0096] Preferably, step S24 includes the following steps:

[0097] Step S241: aligning the network information geographic location data, the network information time series data and the network information semantic data to generate a network information alignment data set;

[0098] Step S242: time-position tagging is performed on the network information geographic location data and the network information time series data to generate network time-position path change data; a propagation node graph is designed for the network time-position path change data based on the graph structure, and node attributes are attached using the network information semantic data to generate a network information node structure graph;

[0099] Step S243: Perform community detection processing on the network information node structure diagram and the network information alignment data set to generate network information dissemination structure data.

[0100] In an embodiment of the present invention, the fusion of multi-dimensional data is achieved by synchronously processing geographic location data, time series data and semantic data. This process uses a timestamp-based alignment algorithm to ensure the consistency of data of different dimensions in time and space. Through data alignment, the generated network information alignment data set provides a unified basic data set for subsequent analysis, in which each data point contains synchronized geographic, temporal and semantic information. This alignment process ensures the integrity and consistency of the data, so that subsequent analysis can be carried out under a unified framework. The network information geographic location data and time series data are time-position marked, and the time and location identifiers are added to each data point using the spatiotemporal marking technology. This process is applied in spatiotemporal data analysis, and network time-location path change data is generated by marking the time and location attributes of each data point. The network time-location path change data is propagated node graph design based on the graph structure. This link uses nodes and edges in graph theory to represent the path and dynamic changes of information propagation. In the graph structure design, nodes represent the key positions of information propagation, and edges represent the propagation path of information between these nodes. Through the design of this graph structure, the propagation path and dynamic evolution of information in the network can be clearly displayed. The network information semantic data is attached to the graph structure as node attributes. The generated network information node structure graph can combine spatial, temporal and semantic data to characterize the characteristics of information dissemination in a more comprehensive way. Finally, the network information node structure graph and the network information alignment dataset are processed for community detection. Community detection technology is an important analysis method in graph theory, which is used to discover node clusters in graph structures. By applying community detection algorithms (such as Louvain algorithm and Girvan-Newman algorithm), clusters of nodes with close connections in the network are identified. These clusters represent key communities and subnetworks in information dissemination.

[0101] As an example of the present invention, refer to Figure 3As shown, in this example, step S3 includes:

[0102] Step S31: Evaluate the influence factors of network information geographic location data, network information time series data and network information semantic data based on a preset network information influence model to generate network information node influence data;

[0103] Step S32: using a dynamic propagation probability function to simulate the propagation probability of the network information node influence data to generate a dynamic propagation path;

[0104] Step S33: constructing an information diffusion mechanism based on the network information node influence data and the dynamic propagation path to generate a network information diffusion mechanism;

[0105] Step S34: Map the network information node influence data, dynamic propagation path and network information diffusion mechanism to the network information model to generate a preliminary information propagation model.

[0106] In the embodiment of the present invention, a preset network information influence model is adopted to comprehensively evaluate multi-source data. The network information geographic location data, time series data and semantic data here represent the spatial distribution, time change trend and content characteristics of the information respectively. Through the impact factor evaluation, this step quantifies the potential influence of these data on information dissemination and generates network information node influence data. The key technologies in this process include data normalization processing and weight allocation to ensure the objectivity and accuracy of the impact factor evaluation. The node influence data obtained by the evaluation is simulated by the propagation probability through the dynamic propagation probability function. The dynamic propagation probability function is an adaptive function based on historical propagation data and current node status, which can adjust the probability of information propagation between nodes in real time and simulate a dynamic propagation path that is closer to reality. The simulation of propagation probability not only takes into account the influence of the node itself, but also combines the network structure characteristics and external environmental variables, so that the generated propagation path is more practical. The construction of the information diffusion mechanism relies on the node influence data and dynamic propagation path generated previously. The information diffusion mechanism identifies key propagation paths and potential diffusion bottlenecks by analyzing the propagation mode of network information between different nodes.

[0107] Preferably, step S31 includes the following steps:

[0108] Step S311: evaluating the influence factors of the network information geographic location data, network information time series data and network information semantic data based on a preset network information influence model, and generating a geographic location factor, a time series factor and a semantic information factor;

[0109] Step S312: Mapping the geographical location factor, the time series factor and the semantic information factor for information propagation association to generate a network information factor association analysis result; assigning weighted parameters to the network information factor association analysis result to generate network information factor parameter weighted data;

[0110] Step S313: Use the fuzzy comprehensive evaluation method to divide the membership of the network information factor association analysis results to generate network information membership data; quantify the influence of the network information factor parameter weighted data and the network information membership data to generate network information node influence data.

[0111] In the embodiment of the present invention, based on the preset network information influence model, the influence factor evaluation is performed on the geographic location data, time series data and semantic data. This process generates geographic location factors, time series factors and semantic information factors respectively by weight analysis of different data types. The technical means in this process mainly include the standardization of multidimensional data and the preliminary evaluation of the influence of each factor by weighted average method. The geographic location factor, time series factor and semantic information factor are integrated, and the network information factor association analysis results are generated by information propagation association mapping technology. This mapping process uses the association analysis algorithm to reveal the interaction between the factors, and quantitatively evaluates the association between the factors by statistical correlation analysis and data mining technology. Weighted parameters are assigned to these association analysis results to generate network information factor parameter weighted data. In this step, the technical means include the application of weighted regression model to ensure that the contribution of each factor to the overall propagation mechanism is accurately reflected. The fuzzy comprehensive evaluation method is further used to divide the membership of the network information factor association analysis results to generate network information membership data. The fuzzy comprehensive evaluation method is used here to deal with data with strong uncertainty and fuzziness, especially when multidimensional data interact, to effectively convert fuzzy information into membership, and then classify the propagation characteristics of network information.

[0112] Preferably, step S32 includes the following steps:

[0113] Step S321: Performing propagation path potential analysis on the network information node influence data to generate bright node potential analysis data; performing potential propagation link identification on the node potential analysis data to generate potential node analysis data;

[0114] Step S322: Perform dynamic spatiotemporal mapping comparison on the bright node potential analysis data and the latent node analysis data to generate dynamic propagation mapping data;

[0115] Step S323: Use the dynamic propagation probability function to perform propagation probability simulation on the dynamic propagation mapping data to generate a dynamic propagation path.

[0116] In an embodiment of the present invention, network information node influence data is used for propagation path potential analysis to generate explicit node potential analysis data. The propagation path potential analysis here relies on influence data, and quantifies the potential influence of nodes in information propagation through a comprehensive evaluation of the node's geographical location, time dimension, and semantic features. Technical means include using clustering algorithms to identify high-potential nodes, and combining regression analysis to predict the potential of these nodes in future propagation. Based on the explicit node potential analysis data, potential propagation links are identified to generate potential node analysis data. In this process, the shortest path algorithm and random walk algorithm in graph theory are used to analyze the propagation path between nodes, and potential links that are not covered by explicit propagation paths are identified. By constructing a potential propagation link model, those paths that have not yet been actually propagated but have high propagation potential can be identified, which provides data support for dynamic mapping in subsequent steps. The explicit node potential analysis data and the potential node analysis data are dynamically mapped and compared to generate dynamic propagation mapping data. Dynamic spatiotemporal mapping maps the propagation potential data of each node into a multidimensional space through time series analysis and spatial mapping technology, showing the spatiotemporal variation law of information propagation. Through this mapping, it is possible to monitor changes in node influence in a dynamic environment and update the propagation paths at different times and locations in real time. Key technologies include the construction of dynamic Bayesian networks and spatiotemporal graph structures. Through these technical means, a propagation mapping model that can be updated with time and space changes is generated. The dynamic propagation probability function is used to simulate the propagation probability of the dynamic propagation mapping data to generate a dynamic propagation path. The dynamic propagation probability function is a function based on historical propagation data and the current state of the node. It can calculate the probability of information propagation between nodes based on the spatiotemporal characteristics of the node and its position in the propagation link. Through the combination of Monte Carlo simulation and Markov chain, a variety of propagation paths are simulated to ensure the prediction accuracy of the dynamic propagation path.

[0117] Preferably, step S33 includes the following steps:

[0118] Step S331: classifying the network information node influence data by node role to generate a network information node classification result, wherein the network information node classification result includes core nodes, edge nodes and relay nodes;

[0119] Step S332: construct a robust core for the core node to generate a core preferred link; connect redundant nodes in series on the edge node to generate an edge redundant link; optimize the transmission link on the relay node to generate a relay node link;

[0120] Step S333: Use network flow technology to construct an information diffusion mechanism for core preferred links, edge redundant links, relay node links and dynamic propagation paths to generate a network information diffusion mechanism.

[0121] In an embodiment of the present invention, node role classification is performed based on network information node influence data, and network information node classification results including core nodes, edge nodes and relay nodes are generated. This process utilizes clustering algorithms and influence scoring models, and classifies nodes into different roles by analyzing their status and role in the network. Core nodes are the main hubs for information dissemination, and therefore have a higher network influence; edge nodes are responsible for the transmission of boundary information, and relay nodes play the role of information transfer. Through this classification, the functional positioning of different nodes in the network can be clarified, laying the foundation for subsequent path optimization. Targeted link optimization processing is performed on different types of nodes. Robust core construction is performed on core nodes to generate core preferred links. This process strengthens the connection between core nodes to ensure that the connection of these key nodes in information dissemination can withstand network failures and attacks. Robust construction uses the minimum spanning tree algorithm and redundant link design in graph theory, and through these means, the connection stability between core nodes is optimized. For edge nodes, redundant node series connection technology is used to generate edge redundant links. This improves the reliability and redundancy of information at the network boundary by adding additional redundant paths between edge nodes. Finally, the transmission link of the relay node is optimized to generate the relay node link. The transmission link optimization mainly uses path selection algorithms, such as the Dijkstra algorithm and the traffic balancing algorithm, to ensure that the relay node can select the optimal path during information transmission, thereby reducing delays and information loss. The network flow technology is used to construct the information diffusion mechanism for the core preferred links, edge redundant links, relay node links, and dynamic propagation paths generated above, and finally generate the network information diffusion mechanism. Network flow technology refers to the dynamic process of simulating information flow in the network structure, using the maximum flow minimum cut algorithm and traffic analysis technology to optimize the efficiency of information propagation. By combining the high stability of the core link, the redundancy of the edge link, and the optimization of the relay link, the information diffusion mechanism can effectively balance the network load, prevent the formation of information bottlenecks, and improve the speed and accuracy of overall information propagation.

[0122] Preferably, step S34 includes the following steps:

[0123] Step S341: Establish an inter-node strength matrix for network information node influence data to generate a network information inter-node strength matrix; construct a propagation link weight matrix for the dynamic propagation path to generate a network information propagation link weight matrix; use the information diffusion mechanism as a dynamic weight parameter matrix to generate a network information dynamic weight parameter matrix;

[0124] Step S342: performing multiple matrix products on the network information node strength matrix, the network information transmission link weight matrix, and the network information dynamic weight parameter matrix to generate a network information transmission data matrix;

[0125] Step S343: Use the node influence data as the initial model parameters to generate a model for the network information propagation data matrix to generate a preliminary information propagation model.

[0126] In the embodiment of the present invention, the node strength matrix is ​​established for the network information node influence data to generate the network information node strength matrix. This process calculates the interaction strength between nodes, uses weighted graphs and node degree analysis, and generates a matrix representing the connection strength between each pair of nodes in the network. This matrix reveals the potential influence relationship between nodes, laying a data foundation for the optimization of subsequent propagation paths. Next, the propagation link weight matrix is ​​constructed for the dynamic propagation path to generate a network information propagation link weight matrix. This weight matrix reflects the probability of information transmission and the propagation resistance on different propagation paths, and is optimized by a path calculation algorithm (such as the Bellman-Ford algorithm) to effectively simulate the characteristics of information flow in different propagation scenarios. In addition, the information diffusion mechanism is converted into a dynamic weight parameter matrix to generate a network information dynamic weight parameter matrix. This step ensures that the changes in the propagation environment and node status can be reflected in real time during the information propagation process by matrixing the dynamically changing parameters in the diffusion mechanism. The above-mentioned multiple matrices - the node strength matrix, the propagation link weight matrix, and the dynamic weight parameter matrix - are multiplied to generate a network information propagation data matrix. The process of matrix multiplication is essentially a comprehensive calculation of data of different dimensions. Through the product operation between matrices, the interactions between nodes, path weights, and dynamic parameters can be integrated into a unified data matrix. This data matrix not only contains the basic information of the network structure, but also reflects the various influencing factors in the process of information propagation. It is a comprehensive description of the entire network propagation state. Finally, the node influence data is used as the initial parameter of the model, and the network information propagation data matrix is ​​used for model generation to finally generate a preliminary information propagation model. This step constructs a preliminary information propagation model by combining the previously generated propagation data matrix with the node influence data and applying model training methods (such as gradient descent and iterative optimization).

[0127] As an example of the present invention, refer to Figure 4 As shown, in this example, step S4 includes:

[0128] Step S41: performing multi-information diffusion mode simulation on the preliminary information propagation model to generate network information simulation data, wherein the network information simulation data includes network information pressure simulation data and network information normal flow data;

[0129] Step S42: tracing the flow information pressure source of the network information pressure simulation data to generate pressure simulation test cause data; adjusting the influence factor weight of the network information node influence data based on the pressure simulation test cause data to generate network information influence adjustment data; performing network flow information detection on the network information normal flow data based on a preset threshold to generate network information propagation detection data;

[0130] Step S43: adjusting model parameters of the preliminary information propagation model based on the network information influence adjustment data and the network information propagation detection data, thereby generating a network information propagation model.

[0131] In an embodiment of the present invention, a preliminary information propagation model is simulated in multiple information diffusion modes to generate network information simulation data. The simulation data here includes network information pressure simulation data and network information normal flow data, which are used to simulate the information propagation dynamics of the network under high load and normal conditions respectively. By using multiple information diffusion mode simulation technology, various information propagation scenarios can be simulated in a virtual environment, including different network topologies, information diffusion rates and propagation paths, thereby laying a data foundation for subsequent pressure analysis and flow detection. The network information pressure simulation data is traced for flow information pressure to generate pressure simulation test cause data. This step analyzes the abnormal flow and propagation bottlenecks shown in the simulation data and uses a tracing algorithm (such as a tracing method based on back propagation) to determine the key factors that cause network pressure. These factors include overload of network nodes, insufficient link bandwidth, and propagation barriers of certain specific information content. The generated pressure simulation test cause data provides a basis for the subsequent adjustment of influence factor weights. Based on these data, the influence factor weights of network information node influence data are adjusted to generate network information influence adjustment data. This adjustment process enables the model to more accurately reflect the actual propagation dynamics by redistributing weights to the influence of nodes. The adjusted influence data can reflect the actual contribution of each node under stress conditions, thereby optimizing the information propagation path and speed. Network traffic information detection is also performed on the normal network information traffic data based on the preset threshold to generate network information propagation detection data. Traffic detection identifies potential problems in the normal propagation process by setting key indicators (such as packet loss rate and transmission delay). These data are used to further calibrate the model to ensure that the model can effectively cope with different network conditions. Based on the network information influence adjustment data and the network information propagation detection data, the model parameters of the preliminary information propagation model are adjusted to generate a network information propagation model. The core of this step is to dynamically adjust the model parameters (such as propagation rate, node influence, and path selection priority) so that the model can better fit the actual network environment. Parameter adjustment is achieved through iterative optimization algorithms (such as gradient descent and Bayesian optimization). In each iteration, the model is fine-tuned based on the simulation data and detection results of the previous round until the model can stably predict the propagation behavior of information in the network.

[0132] The beneficial effect of the present invention is that by acquiring and preprocessing time dimension data, space dimension data and content images, a network information propagation data set with spatiotemporal dynamics and content characteristics is generated. This process effectively integrates data of different dimensions and lays the foundation for the construction of subsequent information propagation models. Through geographic location clustering analysis, time series segmentation processing and semantic content feature extraction, the complex information in the original data is hierarchically structured, and geographic location data, time series data and semantic data are generated respectively, which not only ensures the consistency and integrity of the data in space, time and content, but also improves the availability and processing efficiency of the data. Using the propagation structure construction step, the above-mentioned dimensional data are fused to generate network information propagation structure data. This step reveals the inherent laws and structural characteristics of information propagation through cross-analysis of geographic, time and semantic dimensions, and lays a solid foundation for subsequent impact factor evaluation. In the impact factor evaluation link, the multidimensional data is quantitatively analyzed through the preset network information influence model to generate network information node influence data, thereby revealing the key role of each node in the propagation process. The propagation probability simulation of node influence data using dynamic propagation probability function not only reflects the dynamic characteristics of information propagation, but also provides the ability to predict the actual propagation path through the construction of a preliminary information propagation model. Finally, the final network information propagation model is generated through simulation and parameter adjustment of the preliminary information propagation model. The core of this step is to adjust the feedback based on the simulation data, so that the model has higher accuracy and adaptability, and can better reflect the complex changes in the actual propagation process. Therefore, the present invention solves the problem of insufficient stability of traditional information propagation models under various network conditions through refined node classification and dynamic optimization mechanism, and improves the accuracy of the model in practical applications.

[0133] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0134] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A modeling method for a network information propagation model, characterized in that: The following steps are involved: Step S1: obtaining time dimension data, space dimension data and content image; preprocessing the time dimension data, space dimension data and content image to generate a network information propagation data set, wherein the network information propagation data set includes network information time dimension data, network information space dimension data and network information content dimension data; Step S2: Performing geographic location cluster analysis on the network information spatial dimension data to generate network information geographic location data; Perform time series segmentation processing on the network information time dimension data to generate network information time series data; Extract semantic content features from network information content dimension data to generate network information semantic data; construct propagation structure from network information geographic location data, network information time series data and network information semantic data to generate network information propagation structure data; Step S3: Based on the preset network information influence model, the network information geographic location data, network information time series data and network information semantic data are evaluated for influence factors to generate network information node influence data; the network information node influence data is simulated for propagation probability using the dynamic propagation probability function to generate a dynamic propagation path; the network information node influence data and the dynamic propagation path are mapped to the network information model to generate a preliminary information propagation model; Step S4: simulating the information diffusion mode of the preliminary information propagation model to generate network information simulation data; The model parameters of the preliminary information propagation model are adjusted based on the network information simulation data to generate a network information propagation model.

2. The modeling method of the network information propagation model according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Acquire time dimension data, space dimension data and content picture data; Step S12: extracting content data from the content image using OCR technology to generate content dimension data; Step S13: performing time data discretization processing on the time dimension data to generate network information time dimension data; performing coordinate system one processing on the space dimension data to generate network information space dimension data; Step S14: performing text cleaning on the content dimension data to generate content text cleaning data; performing keyword extraction on the content text cleaning data to generate text content keyword data; performing text vectorization processing on the text content keyword data to generate network information content dimension data; Step S15: Merge the network information time dimension data, the network information space dimension data and the network information content dimension data to generate a network information propagation data set.

3. The modeling method of the network information propagation model according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: Identify information clusters on the network information spatial dimension data to generate network information cluster data; locate hotspots on the network information cluster data to generate network information hotspot location data; cluster geographical locations on the network information hotspot location data to generate network information geographical location data; Step S22: performing periodic time series recognition on the network information time dimension data to generate network information periodic data; performing time series segmentation processing on the network information periodic data to generate network information time series data; Step S23: Calculate the word weights of the network information content dimension data to generate network information word weight data; perform semantic feature extraction processing on the network information word weight data to generate network information semantic feature data; perform semantic data generation on the network information semantic feature data to generate network information semantic data; Step S24: constructing a propagation structure for the network information geographic location data, the network information time series data and the network information semantic data to generate network information propagation structure data.

4. The modeling method of the network information propagation model according to claim 3 is characterized in that: Step S24 includes the following steps: Step S241: aligning the network information geographic location data, the network information time series data and the network information semantic data to generate a network information alignment data set; Step S242: time-position tagging is performed on the network information geographic location data and the network information time series data to generate network time-position path change data; a propagation node graph is designed for the network time-position path change data based on the graph structure, and node attributes are attached using the network information semantic data to generate a network information node structure graph; Step S243: Perform community detection processing on the network information node structure diagram and the network information alignment data set to generate network information dissemination structure data.

5. The modeling method of the network information propagation model according to claim 1 is characterized in that: Step S3 includes the following steps: Step S31: Evaluate the influence factors of network information geographic location data, network information time series data and network information semantic data based on a preset network information influence model to generate network information node influence data; Step S32: using a dynamic propagation probability function to simulate the propagation probability of the network information node influence data to generate a dynamic propagation path; Step S33: constructing an information diffusion mechanism based on the network information node influence data and the dynamic propagation path to generate a network information diffusion mechanism; Step S34: Map the network information node influence data, dynamic propagation path and network information diffusion mechanism to the network information model to generate a preliminary information propagation model.

6. The modeling method of the network information propagation model according to claim 5 is characterized in that: Step S31 includes the following steps: Step S311: evaluating the influence factors of the network information geographic location data, network information time series data and network information semantic data based on a preset network information influence model, and generating a geographic location factor, a time series factor and a semantic information factor; Step S312: Mapping the geographical location factor, the time series factor and the semantic information factor for information propagation association to generate a network information factor association analysis result; assigning weighted parameters to the network information factor association analysis result to generate network information factor parameter weighted data; Step S313: Use the fuzzy comprehensive evaluation method to divide the membership of the network information factor association analysis results to generate network information membership data; quantify the influence of the network information factor parameter weighted data and the network information membership data to generate network information node influence data.

7. The modeling method of network information propagation model according to claim 5, characterized in that: Step S32 includes the following steps: Step S321: Performing propagation path potential analysis on the network information node influence data to generate bright node potential analysis data; performing potential propagation link identification on the node potential analysis data to generate potential node analysis data; Step S322: Perform dynamic spatiotemporal mapping comparison on the bright node potential analysis data and the latent node analysis data to generate dynamic propagation mapping data; Step S323: Use the dynamic propagation probability function to perform propagation probability simulation on the dynamic propagation mapping data to generate a dynamic propagation path.

8. The modeling method of network information propagation model according to claim 5, characterized in that: Step S33 includes the following steps: Step S331: classifying the network information node influence data by node role to generate a network information node classification result, wherein the network information node classification result includes core nodes, edge nodes and relay nodes; Step S332: construct a robust core for the core node to generate a core preferred link; connect redundant nodes in series on the edge node to generate an edge redundant link; optimize the transmission link on the relay node to generate a relay node link; Step S333: Use network flow technology to construct an information diffusion mechanism for core preferred links, edge redundant links, relay node links and dynamic propagation paths to generate a network information diffusion mechanism.

9. The modeling method of network information propagation model according to claim 5, characterized in that: Step S34 includes the following steps: Step S341: Establish an inter-node strength matrix for network information node influence data to generate a network information inter-node strength matrix; construct a propagation link weight matrix for the dynamic propagation path to generate a network information propagation link weight matrix; use the information diffusion mechanism as a dynamic weight parameter matrix to generate a network information dynamic weight parameter matrix; Step S342: performing multiple matrix products on the network information node strength matrix, the network information transmission link weight matrix, and the network information dynamic weight parameter matrix to generate a network information transmission data matrix; Step S343: Use the node influence data as the initial model parameters to generate a model for the network information propagation data matrix to generate a preliminary information propagation model.

10. The modeling method of network information propagation model according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: performing multi-information diffusion mode simulation on the preliminary information propagation model to generate network information simulation data, wherein the network information simulation data includes network information pressure simulation data and network information normal flow data; Step S42: tracing the flow information pressure source of the network information pressure simulation data to generate pressure simulation test cause data; adjusting the influence factor weight of the network information node influence data based on the pressure simulation test cause data to generate network information influence adjustment data; performing network flow information detection on the network information normal flow data based on a preset threshold to generate network information propagation detection data; Step S43: adjusting the model parameters of the preliminary information propagation model based on the network information influence adjustment data and the network information propagation detection data, thereby generating a network information propagation model.

Images