Financial behavior data security analysis method and system based on multi-source data

Abstract

The invention relates to the technical field of financial data analysis, in particular to a financial behavior data security analysis method and system based on multi-source data, and the method comprises the following steps: S1, data collection and integration: collecting financial behavior data; s2, relation reasoning based on a graph: constructing a knowledge graph, analyzing financial behavior data in the knowledge graph, and identifying a potential risk relation and a behavior mode; s3, real-time monitoring and feedback: the financial behaviors of the user are monitored in real time; s4, a self-learning mechanism: forming the self-learning mechanism according to feedback information monitored in real time so as to adapt to a new financial environment and user behavior changes; and S5, risk report generation: generating a risk report according to an analysis result of the financial behavior data in the knowledge graph. According to the invention, newest financial behavior changes can be reflected in time, and rapid and effective risk early warning is provided for managers.

Description

Technical Field

[0001] The present invention relates to the technical field of financial data analysis, and in particular to a method and system for secure analysis of financial behavior data based on multi-source data. Background Art

[0002] With the digitization and complexity of global financial markets, the volume and variety of financial behavior data are increasing. Financial institutions not only need to process internal data from transaction records, user behavior logs, etc., but also need to integrate external data sources, such as social interaction information. This fusion of multi-source data helps to fully understand users' financial behavior patterns and identify potential financial risks in a timely manner. However, the core challenge in financial risk management lies in how to efficiently and accurately analyze these massive multi-source data, discover hidden risk relationships, and protect the security and stability of the financial system.

[0003] Existing financial risk identification methods usually rely on a single data source or static analysis model and are unable to effectively integrate and dynamically update information from multiple data sources. The limitation of this method is that it has a weak ability to capture complex user behaviors and cross-source data, and it is difficult to identify potential abnormal behavior patterns. In addition, many systems lack self-learning capabilities and cannot automatically adjust model parameters according to changes in the financial environment, resulting in a slow response when facing emerging risks and an inability to provide timely risk warnings and decision support. These defects limit the actual application effect of existing technologies in financial risk management.

[0004] The purpose of the present invention is to provide a financial behavior data security analysis method and system based on multi-source data, to help decision makers grasp the risk dynamics in a timely manner and take corresponding measures, and to significantly improve the efficiency, accuracy and intelligence level of financial data security analysis. Summary of the invention

[0005] The present invention provides a method and system for security analysis of financial behavior data based on multi-source data.

[0006] A financial behavior data security analysis method based on multi-source data includes the following steps:

[0007] S1, data collection and integration: collect financial behavior data from various sources, and pre-process and integrate the collected financial behavior data;

[0008] S2, graph-based relational reasoning: construct a knowledge graph, define users, transactions, events and their relationships, analyze the financial behavior data in the knowledge graph, and identify potential risk relationships and behavior patterns;

[0009] S3, real-time monitoring and feedback: real-time monitoring of users' financial behaviors, dynamic update of newly generated financial behavior data to the knowledge graph, feedback to users and managers based on real-time analysis results, and provide risk warnings;

[0010] S4, self-learning mechanism: Based on the feedback information from real-time monitoring, a self-learning mechanism is formed to adapt to the new financial environment and changes in user behavior;

[0011] S5, risk report generation: Generate a risk report based on the analysis results of the financial behavior data in the knowledge graph, including a description of potential risks, risk level assessment, and corresponding response strategy recommendations.

[0012] Optionally, the data collection and integration in S1 includes:

[0013] S11, data collection: collect financial behavior data from financial transaction records, social interaction information, and user behavior logs;

[0014] S12, data preprocessing: cleaning the collected financial behavior data, removing duplicates and missing values;

[0015] S13, data integration: Integrate the pre-processed financial behavior data, build a unified data format, and merge the data through aggregation algorithms.

[0016] Optionally, the graph-based relational reasoning in S2 includes:

[0017] S21, knowledge graph construction: construct a knowledge graph through the integrated financial behavior data, define nodes (such as users, transactions, events, etc.) and the relationships between them (such as the relationship between users and transactions, and the association between transactions and events);

[0018] S22, Relational Reasoning: Use the relational reasoning algorithm to analyze the constructed knowledge graph and mine potential risk relationships and behavior patterns through node features and adjacency relationships;

[0019] S23, risk relationship identification and visualization: After completing the relationship reasoning, risk relationships are identified and the identification results are converted into visual diagrams.

[0020] Optionally, the knowledge graph construction in S21 includes:

[0021] S211, Node and Relationship Definition: Based on the integrated financial behavior data, define the node types, including users, transactions, and events, and calculate the importance (centrality) of the nodes;

[0022] S212, relationship building: determine the relationship between nodes, including the relationship between users and transactions (such as "initiating a transaction"), the association between transactions and events (such as "affected by market fluctuations"), and calculate the relationship strength.

[0023] Optionally, the relational reasoning algorithm in S22 adopts a graph convolutional network (GCN) algorithm, and the graph convolutional network (GCN) algorithm includes:

[0024] S221, build an adaptive adjacency matrix: In financial scenarios, the relationship between users and transactions changes with time and behavior. To reflect the real-time relationship between users and transactions;

[0025] S222, feature update mechanism: enhance node features by introducing historical behavior features, consider the user's past transaction behavior, and design the feature matrix H 0 Initialize;

[0026] S223, feature aggregation of graph convolutional layer: by introducing weighted aggregation strategy, we can enhance the focus on nodes and combine the importance of users and transactions;

[0027] S224, risk score calculation: Calculate the risk score by combining node features and edge weights.

[0028] Optionally, the risk relationship identification and visualization in S23 includes:

[0029] S231, Risk Relationship Identification: After completing the relationship reasoning based on the graph convolutional network (GCN) algorithm, use node features and edge weights to analyze the potential risk relationships in the financial behavior data, and identify high-risk nodes with abnormal behavior and their associated transactions or users by calculating the risk score of each node;

[0030] S232, Visual diagram generation: The identified high-risk nodes and their related transaction or behavior patterns are converted into visual diagrams, and graphical tools are used to display nodes, edges, and risk relationships between nodes. High-risk nodes are highlighted, and the thickness and color of the edges are used to indicate the strength of the risk association between nodes.

[0031] Optionally, the real-time monitoring and feedback in S3 includes:

[0032] S31, Real-time financial behavior monitoring: Real-time monitoring of users’ financial behavior is carried out through the constructed knowledge graph, and new financial transaction data, user behavior logs and social interaction information are continuously collected;

[0033] S32, dynamically updating the knowledge graph: preprocessing and integrating newly collected financial behavior data, dynamically updating it to the knowledge graph, and adjusting the characteristics and weights of nodes and edges;

[0034] S33, Risk Warning and Feedback: Based on the results of real-time risk identification, risk warnings are automatically generated. When high-risk nodes or potential abnormal behaviors are identified, warning information is immediately sent to users and managers.

[0035] Optionally, the self-learning mechanism in S4 includes:

[0036] S41, feedback information collection: based on the feedback information from real-time monitoring, collect changes in user behavior, dynamics of the financial market, and responses to risk warnings;

[0037] S42, self-learning update: Based on the collected feedback information, the self-learning algorithm is used to dynamically update the knowledge graph and the graph convolutional network (GCN) algorithm. By minimizing the loss function of the prediction error, the graph convolutional network (GCN) algorithm gradually adapts to the new financial environment and user behavior changes;

[0038] S43, weight adjustment and node feature update: During the self-learning update process, the node features and edge weights in the knowledge graph are updated.

[0039] Optionally, the risk report generation in S5 includes:

[0040] S51, risk data analysis result processing: based on the financial behavior data results inferred and analyzed by the graph convolutional network (GCN) algorithm in the knowledge graph, extract the identified potential high-risk nodes and their associations, analyze the abnormalities of transaction patterns and user behaviors, and form description information of potential risks;

[0041] S52, risk level assessment: quantify the identified potential risks, and use a risk scoring algorithm to generate risk levels based on the risk scores of nodes, the behavior patterns of associated nodes, and historical financial behavior data, including low risk (risk score of 0%-40%), medium risk (risk score of 40%-70%), and high risk (risk score of 70%-100%);

[0042] S53, Response strategy recommendations: Based on the results of the risk level assessment, generate corresponding response strategy recommendations, take monitoring measures for low risks, including enhanced transaction monitoring, sending early warning notifications, and account activity records; take preventive measures for medium risks, including triggering multi-factor authentication, temporarily freezing suspicious transactions, restricting account permissions, and increasing manual review; take control and remedial measures for high risks, including account freezing, fund tracking and recovery, initiating emergency notification procedures, and shutting down or restricting trading functions;

[0043] S54, report generation and output: Integrate potential risk descriptions, risk level assessments and response strategies to generate risk reports, and automatically output them to managers or relevant decision makers.

[0044] A financial behavior data security analysis system based on multi-source data, used to implement the above-mentioned financial behavior data security analysis method based on multi-source data, includes the following modules:

[0045] Data collection and integration module: collect financial behavior data from various data sources, and pre-process and integrate the collected financial behavior data;

[0046] Graph-based relational reasoning module: Build a knowledge graph of financial behavior data, define users, transactions, events and their relationships, and analyze the financial behavior data in the knowledge graph to identify potential risk relationships and abnormal behavior patterns;

[0047] Real-time monitoring and feedback module: monitor users’ financial behaviors in real time, dynamically update newly generated financial behavior data to the knowledge graph, and generate risk warnings based on analysis results, which are provided to users and managers;

[0048] Self-learning mechanism module: Based on the feedback information from real-time monitoring, a self-learning mechanism is formed to adapt to the new financial environment and changes in user behavior;

[0049] Risk report generation module: Generates risk reports based on the analysis results of financial behavior data in the knowledge graph, including descriptions of potential risks, risk level assessments, and corresponding response strategy recommendations.

[0050] Beneficial effects of the present invention:

[0051] The present invention, by integrating financial behavior data from multiple data sources, combining knowledge graphs and graph convolutional network algorithms, can efficiently and accurately identify potential financial risk relationships and complex user behavior patterns. The system adopts a real-time monitoring and feedback mechanism and dynamically updates the knowledge graph to ensure that the latest financial behavior changes can be reflected in a timely manner, providing managers with fast and effective risk warnings, thereby improving the real-time and accuracy of financial behavior data analysis and meeting the high requirements of the modern financial market for dynamic risk monitoring.

[0052] The present invention, through a self-learning mechanism combined with real-time feedback information, can continuously optimize system parameters, node characteristics and relationship weights, ensuring that the system can adapt to the ever-changing financial environment and user behavior patterns. By introducing self-learning algorithms and loss functions, the system gradually improves its intelligence level in long-term operation, achieves rapid response and accurate identification of new risks, and greatly improves the efficiency and adaptability of financial data security analysis.

[0053] The present invention, through detailed risk reports, covers the description of potential risks, quantitative risk level assessment and targeted response strategy recommendations, to help financial institutions better understand and manage complex risk environments. By segmenting risk levels and classifying response measures, the system improves decision-making efficiency, enhances managers' ability to cope with risks, and provides financial institutions with forward-looking and intelligent risk management decision support, thereby ensuring the security and reliability of financial services. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0055] Figure 1 This is a schematic diagram of the analysis method flow of an embodiment of the present invention;

[0056] Figure 2 Schematic diagram of system function modules according to an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0058] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0059] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0060] like Figure 1 As shown, a financial behavior data security analysis method based on multi-source data includes the following steps:

[0061] S1, data collection and integration: collect financial behavior data from various sources, and pre-process and integrate the collected financial behavior data;

[0062] S2, graph-based relational reasoning: construct a knowledge graph, define users, transactions, events and their relationships, analyze the financial behavior data in the knowledge graph, and identify potential risk relationships and behavior patterns;

[0063] S3, real-time monitoring and feedback: real-time monitoring of users' financial behaviors, dynamic update of newly generated financial behavior data to the knowledge graph, feedback to users and managers based on real-time analysis results, and provide risk warnings;

[0064] S4, self-learning mechanism: Based on the feedback information from real-time monitoring, a self-learning mechanism is formed to adapt to the new financial environment and changes in user behavior;

[0065] S5, risk report generation: Generate a risk report based on the analysis results of the financial behavior data in the knowledge graph, including a description of potential risks, risk level assessment, and corresponding response strategy recommendations;

[0066] Through the above content, accurate identification and real-time monitoring of potential financial risks are achieved. This method has a self-learning mechanism, which can dynamically adapt to the ever-changing financial environment and user behavior, generate detailed risk reports, and provide effective decision-making support, thereby improving the accuracy and security of financial behavior data analysis and meeting the high requirements of modern financial markets for risk management.

[0067] Data collection and integration in S1 include:

[0068] S11, data collection: collect financial behavior data from financial transaction records, social interaction information, and user behavior logs;

[0069] S12, data preprocessing: clean the collected financial behavior data, remove duplicates and missing values, expressed as:

[0070]

[0071] Among them, X is the observed value, μ is the sample mean, σ is the standard deviation, and when the Z value exceeds the set threshold (such as ±3), it is determined to be an outlier and processed;

[0072] S13, data integration: Integrate the pre-processed financial behavior data, build a unified data format, and merge the data through aggregation algorithms, expressed as:

[0073]

[0074] Among them, W i is the weight of data source i, D i is the importance index of the data source, and n is the total number of data sources;

[0075] Through the above content, not only the quality and consistency of data are improved, but also it helps to identify complex risk patterns and potential abnormal behaviors, thereby enhancing the effectiveness of risk identification. In addition, real-time data updates and preprocessing mechanisms enable financial institutions to respond quickly to market changes and user behaviors, optimize risk management strategies, and improve the security and intelligence of overall financial services.

[0076] Graph-based relational reasoning in S2 includes:

[0077] S21, knowledge graph construction: construct a knowledge graph through the integrated financial behavior data, define nodes (such as users, transactions, events, etc.) and the relationships between them (such as the relationship between users and transactions, and the association between transactions and events);

[0078] S22, Relational Reasoning: Use the relational reasoning algorithm to analyze the constructed knowledge graph and mine potential risk relationships and behavior patterns through node features and adjacency relationships;

[0079] S23, risk relationship identification and visualization: After completing the relationship reasoning, the risk relationship is identified and the identification results are converted into visualization diagrams;

[0080] Through the above content, financial institutions can efficiently explore complex user behaviors and transaction patterns and detect anomalies in a timely manner. In addition, the visual display of risk relationships enables managers to more intuitively understand and respond to potential risks, thereby improving the efficiency and accuracy of risk management decisions and ensuring the security and reliability of financial services.

[0081] The knowledge graph construction in S21 includes:

[0082] S211, Node and Relationship Definition: Based on the integrated financial behavior data, define the node types, including users, transactions, and events, and calculate the importance (centrality) of the nodes, expressed as:

[0083]

[0084] Among them, C(u) is the centrality of node u, N(u) is the set of adjacent nodes connected to node u, and C(v) represents the centrality value of node v;

[0085] S212, relationship construction: determine the relationship between nodes, including the relationship between users and transactions (such as "initiating transactions"), the association between transactions and events (such as "affected by market fluctuations"), and calculate the relationship strength, expressed as:

[0086]

[0087] Where R(u,t) represents the strength of the relationship between user u and transaction t, F(u,t) is the number of transactions initiated by the user, and T(u) is the total number of transactions initiated by the user;

[0088] Through the above content, key factors in financial behavior data can be effectively identified and quantified, which not only enhances the understanding of the complex relationships between users, transactions and events, but also provides a reliable data foundation for subsequent risk identification and analysis. By focusing on high-centrality nodes, financial institutions can more accurately understand potential risks, optimize decision-making processes, and improve overall risk management capabilities, thereby enhancing the security and effectiveness of financial services.

[0089] The relational reasoning algorithm in S22 uses the graph convolutional network (GCN) algorithm, which includes:

[0090] S221, build an adaptive adjacency matrix: In financial scenarios, the relationship between users and transactions changes with time and behavior. To reflect the real-time relationship between users and transactions, it is expressed as:

[0091]

[0092] Among them, A ij is the original adjacency matrix, D i and D j are the degrees of nodes i and j, respectively, which are small constants used to avoid division by zero;

[0093] S222, feature update mechanism: enhance node features by introducing historical behavior features, consider the user's past transaction behavior, and design the feature matrix H 0 Initialize, expressed as:

[0094]

[0095] Among them, f current and They represent the extraction functions of current features and historical features respectively, and τ is the historical time window;

[0096] f current (X i )=W current X i +b current ;

[0097] Among them, X i is the original feature matrix of node i, W current is the weight matrix of the current feature, b current is the bias term;

[0098]

[0099] Among them, X i,t is the feature of node i at time t, α t is the weight associated with time t, reflecting the importance of features at different time points;

[0100] S223, feature aggregation of graph convolution layer: By introducing a weighted aggregation strategy, we can enhance the focus on nodes and combine the importance of users and transactions, expressed as:

[0101]

[0102] Among them, α is the weighting coefficient used to enhance the contribution of self-node features, W (l) is the weight matrix of the lth layer, H (l+1) is the feature matrix of the l+1th layer node, that is, the node feature matrix updated by graph convolution, σ is the ReLU activation function, H (l) is the node feature matrix of the lth layer, representing the node features at the current layer;

[0103] S224, risk score calculation: By combining node features and edge weights, the risk score is calculated, which is expressed as:

[0104]

[0105] Among them, RiskScore(u) is the risk score of node u, which means the risk value calculated by node features and edge weights, N(u) is the set of nodes directly connected to user u, and W r is the risk weight vector, is the node feature of the last layer, is the bi-norm of the feature vector of node v, which represents the modulus of the feature and is used to normalize the feature;

[0106] Through the above content, it is possible to capture the changes in node relationships in financial behavior data in real time and fully consider the impact of users' current and historical behaviors. The weighted feature aggregation strategy can more accurately identify key nodes and improve the ability to mine potential risk relationships and complex behavior patterns. The GCN algorithm not only significantly improves the accuracy and real-time performance of risk identification, but also enhances the system's adaptability in a constantly changing financial environment, ensuring the intelligence and efficiency of financial behavior data security analysis.

[0107] Risk relationship identification and visualization in S23 include:

[0108] S231, Risk Relationship Identification: After completing the relationship reasoning based on the graph convolutional network (GCN) algorithm, use node features and edge weights to analyze the potential risk relationships in the financial behavior data, and identify high-risk nodes with abnormal behavior and their associated transactions or users by calculating the risk score of each node;

[0109] S232, Visualization Diagram Generation: The identified high-risk nodes and their related transaction or behavior patterns are converted into visualization diagrams, and the risk relationships between nodes, edges, and nodes are displayed using graphical tools. High-risk nodes are highlighted, and the strength of risk association between nodes is indicated by the thickness and color of the edges.

[0110] Through the above content, not only can potential risk relationships be accurately identified in complex financial behavior data, but the identification results can also be converted into intuitive visual diagrams, making risk patterns and abnormal behaviors easier to understand and analyze. By highlighting high-risk nodes, decision makers can quickly grasp potential threats and improve risk management efficiency. This visual display method helps to simplify complex risk information, improve real-time and accuracy, and enhance the system's early warning and response capabilities.

[0111] Real-time monitoring and feedback in S3 include:

[0112] S31, Real-time financial behavior monitoring: Real-time monitoring of users’ financial behavior is carried out through the constructed knowledge graph, and new financial transaction data, user behavior logs and social interaction information are continuously collected;

[0113] S32, dynamically updating the knowledge graph: preprocessing and integrating newly collected financial behavior data, dynamically updating it to the knowledge graph, and adjusting the characteristics and weights of nodes and edges;

[0114] S33, risk warning and feedback: automatically generate risk warnings based on the results of real-time risk identification, and immediately send warning information to users and managers when high-risk nodes or potential abnormal behaviors are identified;

[0115] Through the above content, high-risk nodes and potential abnormal behaviors can be quickly identified, and risk warnings can be generated immediately to provide timely feedback to users and managers, ensuring that the system can continuously reflect the latest changes in financial data and significantly improve the response speed and accuracy of risk management. Through continuously updated knowledge graphs and accurate risk identification, financial institutions can quickly respond to potential threats and improve the security and stability of overall financial behavior data.

[0116] The self-learning mechanisms in S4 include:

[0117] S41, feedback information collection: based on the feedback information from real-time monitoring, collect changes in user behavior, dynamics of the financial market, and responses to risk warnings;

[0118] S42, self-learning update: Based on the collected feedback information, the self-learning algorithm is used to dynamically update the knowledge graph and the graph convolutional network (GCN) algorithm. By minimizing the loss function of the prediction error, the graph convolutional network (GCN) algorithm gradually adapts to the new financial environment and user behavior changes. The loss function is expressed as:

[0119]

[0120] in, is the predicted output, y i is the actual feedback risk result, and N is the number of data samples;

[0121] S43, weight adjustment and node feature update: In the process of self-learning update, the node features and edge weights in the knowledge graph are updated, so that the graph convolutional network algorithm can more accurately reflect the latest user behavior patterns and financial transaction dynamics;

[0122] Through the above content, it is possible to continuously optimize and adjust its own parameters, node characteristics and edge weights according to real-time feedback information, so as to adapt to the ever-changing financial environment and user behavior patterns, ensure that the system has the ability to dynamically update, and can respond quickly when facing new financial risks or behavioral changes, and maintain high accuracy and adaptability in risk identification. Through self-learning, the system can continuously enhance its intelligence and accuracy in long-term operation, provide more forward-looking risk management and decision-making support, and significantly improve the reliability of financial data security analysis.

[0123] Risk report generation in S5 includes:

[0124] S51, risk data analysis result processing: based on the financial behavior data results inferred and analyzed by the graph convolutional network (GCN) algorithm in the knowledge graph, extract the identified potential high-risk nodes and their associations, analyze the abnormalities of transaction patterns and user behaviors, and form description information of potential risks;

[0125] S52, risk level assessment: quantify the identified potential risks, and use a risk scoring algorithm to generate risk levels based on the risk scores of nodes, the behavior patterns of associated nodes, and historical financial behavior data, including low risk (risk score of 0%-40%), medium risk (risk score of 40%-70%), and high risk (risk score of 70%-100%), expressed as:

[0126]

[0127] Among them, RiskScore(u) is the risk score of node u, and T is the upper limit of the risk score set by the system, which is used to normalize the risk level and generate a percentage risk level assessment;

[0128] S53, Response strategy recommendations: Based on the results of the risk level assessment, generate corresponding response strategy recommendations, take monitoring measures for low risks, including enhanced transaction monitoring, sending early warning notifications, and account activity records; take preventive measures for medium risks, including triggering multi-factor authentication, temporarily freezing suspicious transactions, restricting account permissions, and increasing manual review; take control and remedial measures for high risks, including account freezing, fund tracking and recovery, initiating emergency notification procedures, and closing or restricting specific transaction functions;

[0129] S54, Report generation and output: Integrate potential risk descriptions, risk level assessments and response strategies to generate risk reports, and automatically output them to managers or relevant decision makers;

[0130] Through the above content, not only can potential risks be described in detail, but also quantitative risk level assessment can help decision makers quickly judge the severity of risks. Further combined with historical data and successful response cases, the system provides targeted risk response suggestions. This reporting method greatly improves the efficiency and accuracy of risk management and provides financial institutions with more forward-looking and intelligent decision-making support.

[0131] like Figure 2 As shown, a financial behavior data security analysis system based on multi-source data is used to implement the above-mentioned financial behavior data security analysis method based on multi-source data, including the following modules:

[0132] Data collection and integration module: collect financial behavior data from various data sources, and pre-process and integrate the collected financial behavior data;

[0133] Graph-based relational reasoning module: Build a knowledge graph of financial behavior data, define users, transactions, events and their relationships, and analyze the financial behavior data in the knowledge graph to identify potential risk relationships and abnormal behavior patterns;

[0134] Real-time monitoring and feedback module: monitor users’ financial behaviors in real time, dynamically update newly generated financial behavior data to the knowledge graph, and generate risk warnings based on analysis results, which are provided to users and managers;

[0135] Self-learning mechanism module: Based on the feedback information from real-time monitoring, a self-learning mechanism is formed to adapt to the new financial environment and changes in user behavior;

[0136] Risk report generation module: Generates risk reports based on the analysis results of financial behavior data in the knowledge graph, including descriptions of potential risks, risk level assessments, and corresponding response strategy recommendations.

[0137] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0138] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A financial behavior data security analysis method based on multi-source data, characterized in that: The following steps are involved: S1, data collection and integration: collect financial behavior data from various sources, and pre-process and integrate the collected financial behavior data; S2, graph-based relational reasoning: construct a knowledge graph, define users, transactions, events and their relationships, analyze the financial behavior data in the knowledge graph, and identify potential risk relationships and behavior patterns; S3, real-time monitoring and feedback: real-time monitoring of users' financial behaviors, dynamic update of newly generated financial behavior data to the knowledge graph, feedback to users and managers based on real-time analysis results, and provide risk warnings; S4, self-learning mechanism: Based on the feedback information from real-time monitoring, a self-learning mechanism is formed to adapt to the new financial environment and changes in user behavior; S5, risk report generation: Generate a risk report based on the analysis results of the financial behavior data in the knowledge graph, including a description of potential risks, risk level assessment, and corresponding response strategy recommendations.

2. According to claim 1, a financial behavior data security analysis method based on multi-source data is characterized in that: The data collection and integration in S1 includes: S11, data collection: collect financial behavior data from financial transaction records, social interaction information, and user behavior logs; S12, data preprocessing: cleaning the collected financial behavior data, removing duplicates and missing values; S13, data integration: Integrate the pre-processed financial behavior data, build a unified data format, and merge the data through aggregation algorithms.

3. The method for security analysis of financial behavior data based on multi-source data according to claim 1, characterized in that: The graph-based relational reasoning in S2 includes: S21, knowledge graph construction: construct a knowledge graph through the integrated financial behavior data, define nodes and their relationships; S22, Relational Reasoning: Use the relational reasoning algorithm to analyze the constructed knowledge graph and mine potential risk relationships and behavior patterns through node features and adjacency relationships; S23, risk relationship identification and visualization: After completing the relationship reasoning, risk relationships are identified and the identification results are converted into visual diagrams.

4. A financial behavior data security analysis method based on multi-source data according to claim 3, characterized in that: The knowledge graph construction in S21 includes: S211, Node and Relationship Definition: Based on the integrated financial behavior data, define node types, including users, transactions, and events, and calculate the importance of nodes; S212, relationship building: determine the relationship between nodes, including the relationship between users and transactions, the association between transactions and events, and calculate the relationship strength.

5. A financial behavior data security analysis method based on multi-source data according to claim 4, characterized in that: The relational reasoning algorithm in S22 adopts a graph convolutional network algorithm, and the graph convolutional network algorithm includes: S221, build an adaptive adjacency matrix: In financial scenarios, the relationship between users and transactions changes with time and behavior. To reflect the real-time relationship between users and transactions; S222, feature update mechanism: enhance node features by introducing historical behavior features, consider the user's past transaction behavior, and design the feature matrix H 0 Initialize; S223, feature aggregation of graph convolutional layer: by introducing weighted aggregation strategy, we can enhance the focus on nodes and combine the importance of users and transactions; S224, risk score calculation: Calculate the risk score by combining node features and edge weights.

6. A financial behavior data security analysis method based on multi-source data according to claim 5, characterized in that: The risk relationship identification and visualization in S23 includes: S231, Risk Relationship Identification: After completing the relationship reasoning based on the graph convolutional network algorithm, use node features and edge weights to analyze the potential risk relationships in the financial behavior data, and identify high-risk nodes with abnormal behavior and their associated transactions or users by calculating the risk score of each node; S232, Visual diagram generation: The identified high-risk nodes and their related transaction or behavior patterns are converted into visual diagrams, and graphical tools are used to display nodes, edges, and risk relationships between nodes. High-risk nodes are highlighted, and the thickness and color of the edges are used to indicate the strength of the risk association between nodes.

7. A method for security analysis of financial behavior data based on multi-source data according to claim 6, characterized in that: The real-time monitoring and feedback in S3 include: S31, Real-time financial behavior monitoring: Real-time monitoring of users’ financial behavior is carried out through the constructed knowledge graph, and new financial transaction data, user behavior logs and social interaction information are continuously collected; S32, dynamically updating the knowledge graph: preprocessing and integrating newly collected financial behavior data, dynamically updating it to the knowledge graph, and adjusting the characteristics and weights of nodes and edges; S33, Risk Warning and Feedback: Based on the results of real-time risk identification, risk warnings are automatically generated. When high-risk nodes or potential abnormal behaviors are identified, warning information is immediately sent to users and managers.

8. A method for security analysis of financial behavior data based on multi-source data according to claim 7, characterized in that: The self-learning mechanism in S4 includes: S41, feedback information collection: based on the feedback information from real-time monitoring, collect changes in user behavior, financial market dynamics, and responses to risk warnings; S42, self-learning update: Based on the collected feedback information, the self-learning algorithm is used to dynamically update the knowledge graph and graph convolutional network algorithm. By minimizing the loss function of the prediction error, the graph convolutional network algorithm gradually adapts to the new financial environment and user behavior changes; S43, weight adjustment and node feature update: During the self-learning update process, the node features and edge weights in the knowledge graph are updated.

9. A method for security analysis of financial behavior data based on multi-source data according to claim 8, characterized in that: The risk report generation in S5 includes: S51, risk data analysis result processing: based on the financial behavior data results in the knowledge graph that have been inferred and analyzed by the graph convolutional network algorithm, the identified potential high-risk nodes and their associations are extracted, the abnormalities of transaction patterns and user behaviors are analyzed, and the description information of potential risks is formed; S52, Risk level assessment: quantify the identified potential risks and use a risk scoring algorithm to generate risk levels, including low risk, medium risk, and high risk, based on the risk score of the node, the behavior pattern of the associated nodes, and historical financial behavior data; S53, Response strategy recommendations: Based on the results of the risk level assessment, generate corresponding response strategy recommendations, take monitoring measures for low risks, including enhanced transaction monitoring, sending early warning notifications, and account activity records; take preventive measures for medium risks, including triggering multi-factor authentication, temporarily freezing suspicious transactions, restricting account permissions, and increasing manual review; take control and remedial measures for high risks, including account freezing, fund tracking and recovery, initiating emergency notification procedures, and shutting down or restricting trading functions; S54, report generation and output: Integrate potential risk descriptions, risk level assessments and response strategies to generate risk reports, and automatically output them to managers or relevant decision makers.

10. A financial behavior data security analysis system based on multi-source data, used to implement a financial behavior data security analysis method based on multi-source data as claimed in any one of claims 1 to 9, characterized in that: Includes the following modules: Data collection and integration module: collect financial behavior data from various data sources, and pre-process and integrate the collected financial behavior data; Graph-based relational reasoning module: Build a knowledge graph of financial behavior data, define users, transactions, events and their relationships, and analyze the financial behavior data in the knowledge graph to identify potential risk relationships and abnormal behavior patterns; Real-time monitoring and feedback module: monitor users’ financial behaviors in real time, dynamically update newly generated financial behavior data to the knowledge graph, and generate risk warnings based on analysis results, which are provided to users and managers; Self-learning mechanism module: Based on the feedback information from real-time monitoring, a self-learning mechanism is formed to adapt to the new financial environment and changes in user behavior; Risk report generation module: Generates risk reports based on the analysis results of financial behavior data in the knowledge graph, including descriptions of potential risks, risk level assessments, and corresponding response strategy recommendations.

Images