Risk element identification and multi-factor association modeling method and system

Abstract

The invention provides a risk factor identification and multi-factor association modeling method and system, and belongs to the field of process industrial risk analysis. The modeling method comprises the following steps: collecting an original text data set, and preprocessing the original text data set to obtain a standardized corpus set; performing risk semantic recognition on the standardized corpus set by utilizing a large language model, and extracting semantic elements; performing semantic normalization processing on the extracted semantic elements, and constructing a risk element coding system; establishing a relation network among risk elements, storing semantic nodes and associated edges into a graph database, and generating a risk event multi-dimensional association model; and carrying out risk analysis in the graph database, and analyzing a propagation path and a key control relationship of a risk event by utilizing a visual interface. According to the method, the identification precision of risk factors is improved, potential risk factors in different data sources are effectively captured, the generalization ability of the model under diversified data sources is improved, and the timeliness and flexibility of a risk management system are ensured.

Description

Technical Field

[0001] This invention relates to the field of risk analysis technology in the process industry, and specifically to a method and system for risk factor identification and multi-factor correlation modeling. Background Technology

[0002] With the increasing complexity of the production environment and the surge in data volume in the process industry, traditional risk management methods face enormous challenges. Existing risk prediction methods mainly rely on statistical analysis of historical data and expert experience, which makes it difficult to predict potential risk events in real time and accurately when facing complex and ever-changing production environments. In particular, the ability to process unstructured data, such as equipment operation reports and accident logs, is insufficient, resulting in low data processing efficiency and a lack of effective correlation modeling between data, making it difficult to support global risk prediction.

[0003] To address the dynamic changes in risk factors in industrial environments and ensure safe and stable production operations, a risk factor identification and multi-factor association modeling method based on large language models and knowledge graphs is needed. Large language models can extract potential risk information from massive amounts of text data and achieve understanding of complex language through deep learning; while knowledge graphs can structurally store and associate risk factors extracted from multiple data sources, providing data support for in-depth risk analysis.

[0004] Existing risk identification technologies suffer from the following main shortcomings: Traditional risk identification methods typically rely on structured data, such as historical accident records and equipment data, but have limited processing capabilities for unstructured text data from different sources and in various formats, making it difficult to effectively extract potential risk factors; Although large language models have made significant progress in the field of natural language processing, in the process industry, existing large language models still struggle to accurately identify specific risk elements in complex industrial data, especially in unstructured text data such as equipment operation reports and accident logs, where the semantic understanding depth of the models remains insufficient; Existing risk prediction models often ignore the complex relationships between different risk elements and cannot effectively establish correlations between multiple dimensions and factors. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for risk factor identification and multi-factor correlation modeling, so as to solve the defects of existing risk identification technologies, such as large heterogeneity of data sources, insufficient semantic understanding, and complex multi-factor correlation modeling.

[0006] To achieve the above objectives, this invention provides a method for risk factor identification and multi-factor association modeling. The modeling method includes: collecting raw text datasets based on operational data from process industry systems, preprocessing the raw text datasets to obtain a standardized corpus; using a large language model to perform risk semantic identification on the standardized corpus, extracting semantic elements, and generating identification results in the form of triples; performing semantic normalization on the extracted semantic elements and constructing a risk factor coding system; based on the identification results in the form of triples, establishing a relationship network between risk elements, and storing semantic nodes and related edges in a graph database to generate a multi-dimensional association model of risk events.

[0007] Optionally, the step of collecting the original text dataset and preprocessing the original text dataset to obtain a standardized corpus includes: collecting operational data from a process industry system to obtain an original text dataset, and performing data cleaning and standardization on the original text dataset; segmenting sentences and semantic units based on long sentences and parallel event descriptions in the original text dataset; splitting multi-entity nested descriptions in the original text dataset using dependency parsing and conjunction matching rules; and uniformly encoding multi-source key fields in the original text dataset to generate a standardized corpus.

[0008] Optionally, the step of using a large language model to perform risk semantic recognition on the standardized corpus and extract semantic elements to generate recognition results in the form of triples includes: designing a knowledge graph structure with risk events as core nodes based on the semantic characteristics of the standardized corpus, predefining node types related to risk events, generating a risk event semantic network, and designing prompt word templates for the large language model; using the large language model to identify text entities and relationships in the standardized corpus and generate recognition results in the form of triples.

[0009] Optionally, the generation of the recognition result in the form of triples includes: based on the recognition result in the form of triples, performing confidence screening; when the confidence score is less than a preset threshold, the corresponding risk relationship is removed to obtain a set of high-confidence risk relationships; converting the set of high-confidence risk relationships into a standard triple format and storing it as a structured data table.

[0010] Optionally, the step of performing semantic normalization on the extracted semantic elements and constructing a risk element coding system includes: based on the extracted semantic elements, performing synonym matching and merging using a thesaurus or a Word2Vec word embedding model; standardizing each relation in the semantic elements to maintain a unified relation type in the knowledge graph; labeling each entity in the semantic elements for unified management by the knowledge graph; and constructing a unified coding system based on each risk element in the semantic elements to integrate multi-source content in the knowledge graph.

[0011] Optionally, establishing a relationship network among risk elements based on the identification results in the form of triples includes: importing the identification results in the form of triples into the Neo4j graph database to construct a relationship network among risk elements; and storing the relationship network among risk elements in the graph database to facilitate subsequent path querying and causal chain tracing.

[0012] Optionally, the modeling method further includes: performing risk analysis in a graph database and using a visualization interface to analyze the propagation path of risk events and key control relationships.

[0013] Optionally, the step of performing risk analysis in a graph database and using a visualization interface to analyze the propagation path and key control relationships of risk events includes: using a clustering algorithm to perform cluster analysis on different risk events to identify potential risk events and their propagation paths; using Neo4jBloom to analyze risk events and their propagation paths, and presenting the risk analysis results to users intuitively through an interactive visualization interface.

[0014] Optionally, the clustering algorithm includes: grouping risk events based on a similarity metric to reveal common characteristics and potential risk patterns among the risk events.

[0015] On the other hand, the present invention provides a risk factor identification and multi-factor association modeling system. The modeling system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the modeling method described in any of the above-mentioned embodiments.

[0016] Through the above technical solutions, this invention, by combining large language models and graph database technology, can efficiently process unstructured text data from multiple data sources and transform it into structured information, significantly improving the accuracy of risk factor identification and solving the problem of insufficient processing capabilities for unstructured data by traditional methods. By using large language models for deep semantic understanding of risk factors, it accurately identifies key elements such as risk events, triggering factors, and control measures in the text and outputs them in the form of triples, effectively capturing potential risk factors from different data sources. Furthermore, by using graph databases for relational modeling, it overcomes the limitations of traditional techniques in modeling relationships between complex risk factors. By uniformly encoding and standardizing risk factors identified from different data sources, it ensures the consistency and comparability of multi-source data. By employing methods such as synonym merging and relation standardization, it solves the problem of data source heterogeneity, ensuring the semantic accuracy of risk factors in the knowledge graph and improving the model's generalization ability under diverse data sources. By constructing a relational network of risk factors through graph databases, and utilizing the efficient query and dynamic update functions provided by graph databases, it can update risk factors and their relationships in real time, ensuring the timeliness and flexibility of the risk management system and providing a more efficient and dynamically updatable risk analysis tool.

[0017] Other features and advantages of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating the risk factor identification and multi-factor correlation modeling method of the present invention; Figure 2 This is a schematic diagram of the preprocessing process for the original text dataset in this invention; Figure 3 This is a flowchart illustrating the process of risk semantic recognition for a standardized corpus in this invention. Figure 4 This is a flowchart illustrating the process of constructing a risk element coding system in this invention; Figure 5 This is a flowchart illustrating the process of establishing a relationship network among risk elements in this invention. Figure 6 This is a schematic diagram of the process of risk analysis in a graph database in this invention; Figure 7 This is the logical relationship diagram of the knowledge graph model in this invention. Detailed Implementation

[0019] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0020] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0021] Please refer to Figure 1 This invention provides a method for risk factor identification and multi-factor correlation modeling, the modeling method including: Step S100: Based on the operating data of the process industry system, collect the original text dataset and preprocess the original text dataset to obtain a standardized corpus.

[0022] Please refer to Figure 2 In this embodiment of the invention, the original text dataset is collected and preprocessed to obtain a standardized corpus, which may include: Step S110: Collect operational data from the process industry system to obtain the raw text dataset, and perform data cleaning and standardization on the raw text dataset.

[0023] In a preferred embodiment of the present invention, the operational data collected from the process industry system may include original risk reports, inspection records and operation logs, etc., which constitute the original text dataset. The original text data is characterized by inconsistent formats, chaotic symbols and noisy information, and needs to be cleaned and standardized. The cleaning process includes removing special symbols, unifying Chinese and English punctuation, and correcting abnormal dates and numerical formats.

[0024] Step S120: Based on the long sentences and parallel event descriptions in the original text dataset, perform sentence segmentation and semantic unit division.

[0025] In a preferred embodiment of the present invention, industrial risk reports in the original text dataset typically contain long sentences and parallel event descriptions. To facilitate subsequent semantic recognition, these reports need to be broken down into the smallest semantic units. Semantic segmentation of the text is performed using a combination of punctuation delimitation and dependency parsing. The system uses a set of punctuation marks as the basis for this segmentation. Automatically recognize sentence boundaries with conjunctions (such as "cause", "due to", "and").

[0026] Step S130: Use dependency parsing and conjunction matching rules to split the multi-entity nested descriptions in the original text dataset.

[0027] In a preferred embodiment of the present invention, industrial risk texts in the original text dataset often contain nested descriptions of multiple entities. The original sentence can be split into multiple single semantic unit sentences by a relation decomposition function. The splitting is automatically completed by dependency parsing and conjunction matching rules to ensure that each sentence contains only one risk relation.

[0028] Step S140: Unify the encoding of multi-source key fields in the original text dataset to generate a standardized corpus.

[0029] In a preferred embodiment of the present invention, in order to eliminate the expression differences caused by different text sources, the key fields in the original text dataset can be uniformly encoded. Various date expressions can be unified into ISO format using the following formula:

[0030] in, This is the original time field. For the standardization results, Year-month-day Clock: minutes: seconds.

[0031] In addition, the location field can be numbered according to the three levels of plant area, region and equipment, and the equipment field is uniformly coded according to the equipment dictionary, so as to form a standardized corpus and provide consistent and standardized input data for risk factor identification.

[0032] Step S200: Using a large language model, risk semantic identification is performed on the standardized corpus, and semantic elements are extracted to generate identification results in the form of triples.

[0033] Please refer to Figure 3 In this embodiment of the invention, a large language model is used to perform risk semantic identification on a standardized corpus, and semantic elements are extracted to generate identification results in the form of triples, which may include: Step S210: Based on the semantic characteristics of the standardized corpus, design a knowledge graph structure with risk events as the core nodes, predefine the node types related to risk events, generate a risk event semantic network, and design prompt word templates for the large language model.

[0034] Combination Figure 7In a preferred embodiment of the present invention, a graph structure with "risk event" as the core node can be designed based on the semantic characteristics of industrial risk reports. The predefined node types can include risk event, risk time, risk location, risk type, risk factor, control measures, prevention measures, etc. The relationship types can cover "cause and effect", "time", "location", "belonging", "cause", etc., thereby forming a complete semantic network of risk events. Then, prompt word templates are designed to guide the large language model to perform target recognition.

[0035] Step S220: Using a large language model, identify text entities and relations in a standardized corpus and generate recognition results in the form of triples.

[0036] In a preferred embodiment of the present invention, the large language model can identify entities and their relationships in text through contextual semantic understanding. The output of the large language model is represented in the form of structured triples, which can be expressed by the following formula. :

[0037] in, These represent the identified head and tail entities, respectively. It represents the semantic relationships between them and generates a confidence score for each relationship to measure the reliability of the recognition results.

[0038] In this embodiment of the invention, generating a recognition result in the form of a triplet may include: Step S201: Based on the identification results in the form of triples, perform confidence screening. When the confidence score is less than the preset threshold, the corresponding risk relationship is removed to obtain a set of high-confidence risk relationships.

[0039] Step S202: Convert the high-confidence risk relationship set into a standard triplet format and store it as a structured data table for subsequent normalization and relationship modeling.

[0040] Step S300: Perform semantic normalization on the extracted semantic elements and construct a risk element coding system.

[0041] Please refer to Figure 4 In this embodiment of the invention, the extracted semantic elements are subjected to semantic normalization processing, and a risk element coding system is constructed, which may include: Step S310: Based on the extracted semantic elements, perform synonym matching and merging using a thesaurus or through the Word2Vec word embedding model.

[0042] In a preferred embodiment of the present invention, synonyms that may exist in semantic elements from different sources can be identified and merged, and synonym matching and merging can be performed by searching a thesaurus or using the Word2Vec word embedding model, thereby ensuring that the same entities in different data sources can be represented uniformly.

[0043] Step S320: Standardize each relation in the semantic elements to keep the relation types in the knowledge graph consistent.

[0044] In a preferred embodiment of the present invention, all extracted relations can be standardized to ensure the use of a uniform relation type in the knowledge graph, and various causal relations in the text can be unified into standard causal relations to ensure the consistency of all relations in the knowledge graph.

[0045] Step S330: Tag each entity in the semantic elements for unified management by the knowledge graph.

[0046] In a preferred embodiment of the present invention, each identified entity in the semantic elements can be tagged for unified management in the graph. For example, these tags can be risk event types, control measure types, etc., to ensure efficient data classification and querying.

[0047] Step S340: Based on each risk element in the semantic elements, construct a unified coding system to enable the integration of multi-source content in the knowledge graph.

[0048] In a preferred embodiment of the present invention, a unified coding system can be constructed for each risk element to ensure that the same entities and relationships in different data sources can be mapped to a unified code. Through the unified coding system, the content of different data sources can be seamlessly integrated in the knowledge graph.

[0049] Step S400: Based on the identification results in the form of triples, establish a relationship network among risk elements, and store semantic nodes and related edges in a graph database to generate a multidimensional association model of risk events.

[0050] Please refer to Figure 5 In this embodiment of the invention, establishing a relationship network among risk factors based on the identification results in the form of triples may include: Step S410: Import the identification results in the form of triples into the Neo4j graph database to construct a relationship network between risk elements. The entities in each triple are stored as nodes and the relationships are stored as edges.

[0051] Step S420: Store the relationship network between risk factors in a graph database to facilitate subsequent path queries and causal chain tracing.

[0052] In a preferred embodiment of the present invention, the constructed risk element relationship network can be stored in a graph database. The efficient query capabilities of the graph database support complex path queries and causal chain tracing, thereby helping to identify potential risk transmission paths. Specific paths can be queried using the Cypher language, for example... .

[0053] Step S500: Perform risk analysis in the graph database and use a visualization interface to analyze the propagation path of risk events and key control relationships.

[0054] Please refer to Figure 6 In this embodiment of the invention, risk analysis is performed in a graph database, and a visual interface is used to analyze the propagation path and key control relationships of risk events. This may include: Step S510: Use clustering algorithms to perform cluster analysis on different risk events in order to identify potential risk events and their propagation paths.

[0055] In a preferred embodiment of the present invention, a clustering algorithm can be used to perform cluster analysis on different risk events to identify potential high-risk areas or time periods. Specifically, the clustering algorithm can group events based on similarity metrics, revealing common characteristics and potential risk patterns. The K-means clustering algorithm can be expressed by the following formula:

[0056] in, For the first Data points, For the first The clustering algorithm minimizes the sum of squared distances between each point and the cluster center.

[0057] Step S520: Use Neo4jBloom to analyze risk events and their propagation paths, and present the risk analysis results (e.g., causal chains between risk events, key control factors, and scope of impact) to users through an interactive visualization interface.

[0058] On the other hand, embodiments of the present invention provide a risk factor identification and multi-factor association modeling system. The modeling system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement any of the above-mentioned modeling methods.

[0059] Accordingly, embodiments of the present invention provide a method and system for risk factor identification and multi-factor association modeling. The modeling method includes: collecting raw text datasets based on the operational data of a process industry system, and preprocessing the raw text datasets to obtain a standardized corpus; using a large language model to perform risk semantic identification on the standardized corpus, extracting semantic elements, and generating identification results in the form of triples; performing semantic normalization processing on the extracted semantic elements and constructing a risk factor coding system; based on the identification results in the form of triples, establishing a relationship network between risk elements, and storing semantic nodes and related edges in a graph database to generate a multi-dimensional association model of risk events. This invention combines large language models and graph database technology to efficiently process unstructured text data from multiple data sources and transform it into structured information, significantly improving the accuracy of risk factor identification and solving the problem of insufficient processing capabilities for unstructured data by traditional methods. Through deep semantic understanding of risk factors using large language models, it accurately identifies key elements such as risk events, triggering factors, and control measures in the text, outputting them in triplet form. This effectively captures potential risk factors from different data sources. Furthermore, graph database-based association modeling overcomes the limitations of traditional techniques in modeling relationships between complex risk factors. By uniformly encoding and standardizing risk factors identified from different data sources, it ensures the consistency and comparability of multi-source data. Methods such as synonym merging and relationship standardization address the heterogeneity of data sources, ensuring the semantic accuracy of risk factors in the knowledge graph and improving the model's generalization ability across diverse data sources. Finally, by constructing a relationship network of risk factors using a graph database, and utilizing the efficient query and dynamic update functions provided by the graph database, it can update risk factors and their relationships in real time, ensuring the timeliness and flexibility of the risk management system and providing a more efficient and dynamically updatable risk analysis tool.

[0060] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0061] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0062] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0063] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0064] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0065] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0066] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0067] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0068] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for risk factor identification and multi-factor correlation modeling, characterized in that, The modeling method includes: Based on the operational data of the process industry system, a raw text dataset is collected, and the raw text dataset is preprocessed to obtain a standardized corpus. Using a large language model, risk semantic identification is performed on the standardized corpus, and semantic elements are extracted to generate identification results in the form of triples; The extracted semantic elements are semantically normalized, and a risk element coding system is constructed. Based on the identification results in the form of triples, a relationship network among risk elements is established, and semantic nodes and related edges are stored in a graph database to generate a multidimensional association model of risk events.

2. The modeling method according to claim 1, characterized in that, The process of collecting the original text dataset and preprocessing it to obtain a standardized corpus includes: The process industry system's operational data is collected to obtain a raw text dataset, which is then cleaned and standardized. Based on long sentences and parallel event descriptions in the original text dataset, sentence segmentation and semantic unit partitioning are performed. Dependency parsing and conjunction matching rules are used to decompose multi-entity nested descriptions in the original text dataset; The key fields from multiple sources in the original text dataset are uniformly encoded to generate a standardized corpus.

3. The modeling method according to claim 1, characterized in that, The process of using a large language model to perform risk semantic identification on the standardized corpus, extracting semantic elements, and generating identification results in the form of triples includes: Based on the semantic characteristics of the standardized corpus, a knowledge graph structure with risk events as the core nodes is designed, and node types related to risk events are predefined to generate a semantic network of risk events. Prompt word templates are also designed for the large language model. Using a large language model, we can identify text entities and relationships in a standardized corpus and generate recognition results in the form of triples.

4. The modeling method according to claim 1, characterized in that, The recognition results in the form of generated triples include: Based on the identification results in the form of triples, confidence screening is performed. When the confidence score is less than a preset threshold, the corresponding risk relationship is removed to obtain a set of high-confidence risk relationships. The set of high-confidence risk relationships is converted into a standard triplet format and stored as a structured data table.

5. The modeling method according to claim 1, characterized in that, The step of performing semantic normalization on the extracted semantic elements and constructing a risk element coding system includes: Based on the extracted semantic elements, synonym matching and merging are performed using a thesaurus or through the Word2Vec word embedding model; Standardize each relation in the semantic elements to keep the relation types in the knowledge graph consistent; Each entity in the semantic elements is tagged for unified management by the knowledge graph. Based on each risk element in the semantic elements, a unified coding system is constructed to enable the integration of multi-source content in the knowledge graph.

6. The modeling method according to claim 1, characterized in that, The identification results based on triplet form establish a relationship network among risk factors, including: The identification results in the form of triples are imported into the Neo4j graph database to construct a relationship network among risk elements; Store the relationship network between risk factors in a graph database to facilitate subsequent path queries and causal chain tracing.

7. The modeling method according to claim 1, characterized in that, The modeling method also includes: Risk analysis is performed in graph databases, and a visualization interface is used to analyze the propagation paths of risk events and key control relationships.

8. The modeling method according to claim 7, characterized in that, The risk analysis is performed in a graph database, and a visualization interface is used to analyze the propagation path of risk events and key control relationships, including: Clustering algorithms are used to perform cluster analysis on different risk events in order to identify potential risk events and their propagation paths; Neo4jBloom is used to analyze risk events and their propagation paths, and the risk analysis results are presented to users intuitively through an interactive visualization interface.

9. The modeling method according to claim 8, characterized in that, The clustering algorithm includes grouping risk events based on a similarity metric to reveal common characteristics and potential risk patterns among the risk events.

10. A risk factor identification and multi-factor correlation modeling system, characterized in that, The modeling system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the modeling method according to any one of claims 1-9.

Images