Financing guarantee risk dynamic monitoring and early warning method and system based on data fusion
By building a business data transfer platform and utilization period-failure network model, integrating multi-source data and combining risk assessment models, the problems of data silos and static warnings in traditional financing guarantees are solved, and the accuracy and timeliness of risk monitoring are improved.
Patent Information
- Application Number
- CN202410936427.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-12
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-07-12
AI Technical Summary
The traditional financing guarantee risk assessment method is based on limited data and empirical judgment, making it difficult to comprehensively and objectively evaluate corporate risks. The monitoring and early warning method is static and cannot monitor risk changes in real time, resulting in insufficient timeliness and accuracy of early warnings, and the data is scattered in different systems and departments. , resulting in information silos, making it difficult to integrate and analyze.
By building a business data transit platform, integrating batch business data from different servers, using pre-trained deadline-failure network models for data failure analysis, extracting financing guarantee risk characteristics, and combining embedded risk assessment models for evaluation. A warning signal is generated when the risk indicator exceeds the preset value.
It has achieved comprehensive integration and analysis of multi-source data, improved the accuracy and reliability of risk assessment, generated early warning signals in a timely manner, and improved the timeliness and accuracy of monitoring and early warnings.
Smart Images

Figure CN118941376A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of financing guarantee risk early warning, and in particular to a financing guarantee risk dynamic monitoring and early warning method and system based on data fusion. Background Art
[0002] Financing guarantee is a guarantee institution established to support small and medium-sized enterprises in obtaining financing. By providing guarantees, it helps enterprises solve financing problems. Therefore, risk monitoring and early warning for financing guarantee are particularly important, and there are still some problems with existing technologies. First, traditional risk assessment methods for financing guarantee are often based on limited data and experience judgments, which makes it difficult to comprehensively and objectively assess the risk status of enterprises, and is prone to misjudgment and omissions; second, traditional monitoring and early warning methods are usually static, unable to monitor risk changes in real time and issue early warnings in a timely manner, resulting in insufficient timeliness and accuracy of early warnings; third, in traditional financing guarantee business, data from various links are often scattered in different systems and departments, and there are information islands, which makes it difficult to integrate and analyze data, affecting business efficiency. Summary of the invention
[0003] This application provides a dynamic monitoring and early warning method for financing guarantee risks based on data fusion, aiming to solve the problem of information islands in traditional financing guarantee business, which makes data difficult to integrate and analyze, and traditional monitoring and early warning methods are usually static and unable to monitor risk changes in real time and issue timely warnings, resulting in technical problems such as insufficient timeliness and accuracy of warnings.
[0004] In view of the above problems, this application provides a method and system for dynamic monitoring and early warning of financing guarantee risks based on data fusion.
[0005] The first aspect disclosed in the present application provides a dynamic monitoring and early warning method for financing guarantee risk based on data fusion, the method comprising: building a business data transfer platform, wherein the business data transfer platform is a transfer platform for business data transmission between a first service end and a second service end; monitoring a batch business data set associated with the quantitative financing guarantee risk from the first service end, and inputting the batch business data set into the business data transfer platform; the business data transfer platform identifies the batch business data set, determines a first type of batch business data set and a second type of batch business data set, wherein the business data transfer platform includes a pre-trained term-failure network model, and the term-failure network model is used to perform data failure analysis on the batch business data set; feature extraction is performed on the first type of batch business data set and the second type of batch business data set respectively, to obtain a first type of risk feature and a second type of risk feature associated with the quantitative financing guarantee risk; the business data transfer platform transmits the first type of risk feature and the second type of risk feature to the second service end, and the second service end uses an embedded financing guarantee risk assessment model to output a financing guarantee risk index; when the financing guarantee risk index is greater than a preset financing guarantee risk index, a first early warning signal is generated.
[0006] The second aspect disclosed in the present application provides a dynamic monitoring and early warning system for financing guarantee risks based on data fusion, and the system is used for the above-mentioned dynamic monitoring and early warning method for financing guarantee risks based on data fusion, and the system includes: a transfer platform building module, the transfer platform building module is used to build a business data transfer platform, wherein the business data transfer platform is a transfer platform for transmitting business data between a first server end and a second server end; a business data input module, the business data input module is used to monitor batch business data sets associated with the quantitative association of financing guarantee risks from the first server end, and input the batch business data sets into the business data transfer platform; a business data identification module, the business data identification module is used for the business data transfer platform to identify the batch business data sets, determine a first type of batch business data set and a second type of batch business data set, and In the embodiment, the business data transfer platform includes a pre-trained deadline-failure network model, which is used to perform data failure analysis on the batch business data set; a feature extraction module, which is used to perform feature extraction on the first type of batch business data set and the second type of batch business data set respectively, to obtain the first type of risk characteristics and the second type of risk characteristics quantitatively associated with the financing guarantee risk; a risk indicator output module, which is used for the business data transfer platform to transmit the first type of risk characteristics and the second type of risk characteristics to the second server, and the second server uses the embedded financing guarantee risk assessment model to output the financing guarantee risk indicator; an early warning signal generation module, which is used to generate a first early warning signal when the financing guarantee risk indicator is greater than the preset financing guarantee risk indicator.
[0007] According to a third aspect disclosed in the present application, a computer device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the following steps when executing the computer program: building a business data transfer platform, wherein the business data transfer platform is a transfer platform for business data transmission between a first server and a second server; monitoring batch business data sets associated with financing guarantee risk quantification from the first server, and inputting the batch business data sets into the business data transfer platform; the business data transfer platform identifies the batch business data sets, determines a first type of batch business data set and a second type of batch business data set, wherein the business data transfer platform includes a pre-trained term-failure network model, and the term-failure network model is used to perform data failure analysis on the batch business data sets; feature extraction is performed on the first type of batch business data set and the second type of batch business data set, respectively, to obtain a first type of risk feature and a second type of risk feature associated with financing guarantee risk quantification; the business data transfer platform transmits the first type of risk feature and the second type of risk feature to the second server, and the second server uses the embedded financing guarantee risk assessment model to output a financing guarantee risk index; when the financing guarantee risk index is greater than the preset financing guarantee risk index, a first warning signal is generated.
[0008] According to a fourth aspect disclosed in the present application, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: building a business data transfer platform, wherein the business data transfer platform is a transfer platform for business data transmission between a first server and a second server; monitoring batch business data sets associated with financing guarantee risk quantification from the first server, and inputting the batch business data sets into the business data transfer platform; the business data transfer platform identifies the batch business data sets, determines a first type of batch business data set and a second type of batch business data set, wherein the business data transfer platform includes a pre-trained term-failure network model, and the term-failure network model is used to perform data failure analysis on the batch business data sets; feature extraction is performed on the first type of batch business data set and the second type of batch business data set, respectively, to obtain a first type of risk feature and a second type of risk feature associated with financing guarantee risk quantification; the business data transfer platform transmits the first type of risk feature and the second type of risk feature to the second server, and the second server uses the embedded financing guarantee risk assessment model to output a financing guarantee risk index; when the financing guarantee risk index is greater than the preset financing guarantee risk index, a first warning signal is generated.
[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0010] By using the business data transfer platform, batch business data sets from different service ends can be integrated together, and the pre-trained deadline-failure network model can be used to analyze data failure, thereby realizing the comprehensive integration and analysis of multi-source data, and improving the accuracy and comprehensiveness of the monitoring results; by extracting features from the first type of batch business data set and the second type of batch business data set, the risk characteristics associated with the financing guarantee risk quantification are obtained, and the risk assessment is carried out in combination with the embedded financing guarantee risk assessment model, which can fully consider the influence of various risk factors and improve the accuracy and reliability of the assessment results; the embedded financing guarantee risk assessment model is used to conduct financing guarantee risk assessment, and the preset financing guarantee risk indicators are set to make judgments. When the risk indicator exceeds the preset value, an early warning signal can be generated in time, which improves the timeliness and accuracy of monitoring and early warning. In summary, the dynamic monitoring and early warning method of financing guarantee risk based on data fusion improves the accuracy, comprehensiveness and timeliness of risk monitoring and early warning by integrating multi-source data, comprehensively analyzing risk factors, and combining real-time monitoring and early warning mechanisms.
[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] Figure 1 A schematic diagram of a process flow of a dynamic monitoring and early warning method for financing guarantee risks based on data fusion provided in an embodiment of the present application;
[0013] Figure 2 A schematic diagram of the structure of a dynamic monitoring and early warning system for financing guarantee risks based on data fusion provided in an embodiment of the present application;
[0014] Figure 3 This is a diagram of the internal structure of a computer device provided in an embodiment of the present application.
[0015] Explanation of the accompanying drawings: transfer platform building module 10, business data input module 20, business data identification module 30, feature extraction module 40, risk indicator output module 50, warning signal generation module 60. DETAILED DESCRIPTION
[0016] The embodiments of the present application provide a dynamic monitoring and early warning method for financing guarantee risks based on data fusion, thereby solving the problem that in traditional financing guarantee business, there are information islands, which makes data difficult to integrate and analyze, and traditional monitoring and early warning methods are usually static and cannot monitor risk changes in real time and issue timely warnings, resulting in technical problems such as insufficient timeliness and accuracy of warnings.
[0017] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the accompanying drawings of the specification. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0018] like Figure 1 As shown, the embodiment of the present application provides a method for dynamic monitoring and early warning of financing guarantee risks based on data fusion, the method comprising:
[0019] Building a business data transfer platform, wherein the business data transfer platform is a transfer platform for business data transmission between the first server and the second server;
[0020] Analyze the business data transmission needs, including the frequency of data transmission, data volume, transmission security requirements, etc. Based on the demand analysis results, design the overall architecture of the business data transfer platform, determine the interface specifications and communication protocols of the first server and the second server, and build a business data transfer platform. The business data transfer platform is an intermediate platform for business data transmission between the first server and the second server, and is used to realize data transfer, forwarding and processing. The first server is the endpoint for generating or collecting business data, and the second server is the endpoint for receiving and processing business data.
[0021] Monitoring a batch of business data sets associated with the quantification of financing guarantee risks from the first service end, and inputting the batch of business data sets into the business data transfer platform;
[0022] The first server regularly monitors batch business data sets related to the quantification of financing guarantee risks through a monitoring system or data collection program. These data sets may include loan information, guarantee information, risk indicators, etc., and transmits these data sets to the business data transfer platform through a network connection for subsequent data processing and analysis.
[0023] The business data transfer platform identifies the batch business data set to determine a first type of batch business data set and a second type of batch business data set, wherein the business data transfer platform includes a pre-trained deadline-failure network model, and the deadline-failure network model is used to perform data failure analysis on the batch business data set;
[0024] The business data transfer platform first loads the pre-trained deadline-failure network model, which is trained on historical data through machine learning and is used to analyze the failure of business data. The batch business data set is input into the model, and the model performs failure analysis on the business data in the data set based on the characteristics in the data set and the pattern of historical data. According to the analysis results, the batch business data set is divided into the first category and the second category, where the first category of data sets contains high-risk or already failed data, and the second category of data sets contains low-risk or not yet failed data.
[0025] Extracting features from the first type of batch business data set and the second type of batch business data set respectively to obtain first type of risk features and second type of risk features quantitatively associated with financing guarantee risks;
[0026] For the first category of batch business data sets and the second category of batch business data sets, the business data transfer platform extracts features associated with the quantification of financing guarantee risks through data analysis and feature engineering technology. During the feature extraction process, it is necessary to ensure that the extracted features are related to the quantification of financing guarantee risks and can reflect the risk level of the business data. Appropriate features can be selected based on domain knowledge and actual conditions, and ensure that these features can play a role in subsequent risk assessments. By way of example, the extracted features include but are not limited to financial indicators, risk ratings, historical defaults, collateral assessments, etc. After the features are extracted, the first category of risk features and the second category of risk features are obtained for subsequent data transmission and risk assessment use.
[0027] The business data transfer platform transmits the first and second risk characteristics to the second server, and the second server outputs the financing guarantee risk index by using the embedded financing guarantee risk assessment model;
[0028] The business data transfer platform transmits the extracted first-category risk features and second-category risk features to the second server through the network. The second server uses the embedded financing guarantee risk assessment model to conduct risk assessment on the received feature data. This financing guarantee risk assessment model is built based on machine learning and is used to predict and assess financing guarantee risks. Based on the processing of feature data, corresponding financing guarantee risk indicators are generated. These indicators include risk scores, risk levels, probabilities, etc., which are used to indicate the degree of financing guarantee risks and provide data support for subsequent risk monitoring and early warning.
[0029] When the financing guarantee risk indicator is greater than a preset financing guarantee risk indicator, a first warning signal is generated.
[0030] According to actual conditions, historical experience, etc., set preset financing guarantee risk indicators as acceptable risk standards. For the obtained financing guarantee risk indicators, compare them with the preset financing guarantee risk indicators. When the financing guarantee risk indicator is greater than the preset financing guarantee risk indicator, a first warning signal is generated. This warning signal can be a notification or alarm information automatically issued by the system. The first warning signal is notified to relevant personnel so that they can take timely action to deal with the financing guarantee risk.
[0031] Furthermore, the business data transfer platform identifies the batch business data set to determine the first type of batch business data set and the second type of batch business data set, and the method includes:
[0032] Decomposing the business data types of the batch business data set to obtain each business data type;
[0033] Perform failure index analysis on each service data type according to the deadline-failure network model, and output failure factors of each service data type;
[0034] The batch business data sets are classified using the failure factors of various business data types as variables to determine a first type of batch business data sets whose failure factors are greater than or equal to a preset failure factor and a second type of batch business data sets whose failure factors are less than the preset failure factor.
[0035] According to the structure and content of batch business data sets, they are decomposed into different business data types. For each decomposed data, analysis is performed based on the data format, field meaning, business rules, etc. to obtain each business data type.
[0036] The business data transfer platform loads a pre-trained term-failure network model, which is trained through a training sample data set. Based on the failure characteristics of the business data types in financing guarantees learned during the training process, the platform performs failure index analysis on each business data type. Specifically, the longer the time, the more impact each type of data has on the credibility, accuracy, and value of the financing guarantee, and the higher the probability of failure. Based on the results of the failure index analysis, the platform outputs corresponding failure factors for each business data type. These failure factors are indicators that reflect the credibility, accuracy, and value of data in financing guarantees, and can be in the form of probability values, scores, etc.
[0037] Prepare batch business data sets to be classified, and use each business data type and its corresponding failure factor as input data, set a preset failure factor threshold, and use it to classify the first and second batch business data sets. This preset value can be set based on domain expert knowledge, historical data analysis, or business needs. For each piece of data, use the failure factor of each business data type corresponding to it as a variable, and classify the data set according to the preset failure factor threshold. Classify the data set with a failure factor greater than or equal to the preset failure factor as the first batch business data set, and classify the data set with a failure factor less than the preset failure factor as the second batch business data set.
[0038] Furthermore, the method further comprises:
[0039] Get the fully connected neural network layer;
[0040] Obtaining a training sample data set, wherein the training sample data set includes batch business sample data, multiple failure penalty factors, and identification information for identifying the failure factors, wherein the multiple failure penalty factors include deadline-data credibility, deadline-data accuracy, and deadline-data value;
[0041] The fully connected neural network layer is trained according to the training sample data set until the model is in a convergence state, and the deadline-failure network model is output.
[0042] A fully connected neural network is a basic artificial neural network structure, in which each neuron is connected to all neurons in the previous layer. Specifically, the structure of the fully connected neural network is determined, including the number of neurons in the input layer, hidden layer, and output layer, as well as the connection method between the layers. The parameters of the neural network, including weights and biases, are initialized. According to the designed structure, a fully connected neural network layer is built, and an activation function is applied to each connection between the hidden layer and the output layer to introduce nonlinear characteristics and improve the expression ability of the network. After the neural network layer is built, the loss function and optimizer are set. The loss function is used to measure the difference between the model's predicted output and the actual label, and the optimizer is used to adjust the model parameters to minimize the loss function, and finally an initialized fully connected neural network layer is obtained.
[0043] Batch business sample data are business data samples used to train the deadline-failure network model. These sample data can be historical data, simulated data or synthetic data, reflecting various situations and changes in different business scenarios.
[0044] Multiple failure penalty factors are parameters used to measure the degree of data failure. These factors include term-data credibility, term-data accuracy and term-data value. They are used to quantify the failure probability of batch business sample data in financing guarantees and serve as input features of the model. The longer the time, the more affected the credibility, accuracy and value of each type of data in financing guarantees, and the higher the probability of failure.
[0045] The identification information of the failure factor is a label used to distinguish different failure penalty factors. In order to train the deadline-failure network model, each sample data needs to be associated with its corresponding failure penalty factor so that the model can learn the impact of each factor on data failure.
[0046] The obtained batch business sample data, multiple failure penalty factors and identification information of the failure factors are integrated to obtain a training sample data set. Such a data set can be used to train a term-failure network model to learn the failure characteristics of business data types in financing guarantees, thereby conducting failure indicator analysis.
[0047] Use the training sample data set to train the constructed fully connected neural network layer. During the training process, adjust the model parameters through the back propagation algorithm and optimizer so that the model can gradually learn the characteristics and patterns of the data. Monitor the performance of the model on the training sample data set, including the change of the loss function, the improvement of accuracy and other indicators, to ensure that the model gradually converges during the training process. Regularly evaluate the performance of the model on the validation set, and adjust the hyperparameters and structure of the model as needed.
[0048] When the model reaches a convergence state during the training process, for example, when the accuracy reaches the preset conditions or the number of iterations reaches the preset number, the trained deadline-failure network model is output. This model contains the characteristics and patterns of business data failure learned from the training sample data set and can be used for subsequent failure indicator analysis tasks.
[0049] Through the above steps, the fully connected neural network model can be trained on the training sample data set until the model converges and outputs the term-failure network model. This model can learn the failure characteristics of business data types in financing guarantees and provide a basis for subsequent failure indicator analysis.
[0050] Furthermore, the second service end uses the embedded financing guarantee risk assessment model to perform risk assessment on the first risk characteristics and the second risk characteristics, and outputs a financing guarantee risk index;
[0051] Among them, the first type of risk characteristics and the second type of risk characteristics include an optimal weight ratio, and the weight of the first type of risk characteristics is less than the weight of the second type of risk characteristics, and risk assessment is performed based on the optimal weight ratio in combination with the first type of risk characteristics and the second type of risk characteristics.
[0052] The second server activates a pre-embedded financing guarantee risk assessment model, which is a trained and optimized machine learning model. It takes the first type of risk characteristics and the second type of risk characteristics as inputs to the model. The financing guarantee risk assessment model performs risk assessment based on the optimal weight ratio by combining the first type of risk characteristics and the second type of risk characteristics.
[0053] Specifically, a weight ratio of the first-category risk feature and the second-category risk feature is set, and the weight ratio reflects the importance of different features in risk assessment, wherein the weight of the first-category risk feature is less than the weight of the second-category risk feature. According to the set weight ratio, the first-category risk feature and the second-category risk feature are weighted, which can be achieved by simply multiplying each feature by the corresponding weight ratio. The weighted first-category risk feature and the second-category risk feature are merged, for example, by connecting them into a vector or matrix as a joint feature vector.
[0054] Using the embedded financing guarantee risk assessment model, the risk assessment of the joint features is carried out. This assessment process specifically considers the weight ratio of different features. According to the results of the risk assessment, the financing guarantee risk indicators are output. These indicators reflect the risk situation after the assessment based on the joint features. This method can more accurately consider the importance of different features, thereby improving the accuracy and practicality of risk assessment.
[0055] Furthermore, the method for obtaining the optimization weight ratio includes:
[0056] Obtain m initial sub-fusion models, and assign m different weight ratios λ1:λ2 to the m initial sub-fusion models, wherein the weight λ1 of the first type of risk feature is less than the weight λ2 of the second type of risk feature;
[0057] Constructing a fusion sample set, the fusion sample set comprising a training sample set of the first type of risk characteristics, a training sample set of the second type of risk characteristics, and a financing guarantee risk verification sample set of the first type of risk characteristics and the second type of risk characteristics;
[0058] The m initial sub-fusion models are trained according to the fusion sample set, and an integrated fusion model is output.
[0059] Obtain m initial sub-fusion models, which are models with randomly initialized parameters, set different weight ratios λ1 and λ2, where λ1 represents the weight of the first type of risk feature, λ2 represents the weight of the second type of risk feature, and λ1<λ2, and assign different weight ratios to each initial sub-fusion model. This can be achieved by adjusting the parameters of each model to reflect different weight ratios.
[0060] Select and extract the sample data required for the first type of risk features from the available data, and construct it into a training sample set to obtain the training sample set of the first type of risk features. These sample data contain the first type of risk features and corresponding labels for model training. The training sample set of the second type of risk features is obtained in the same way.
[0061] A part of the data is selected as the financing guarantee risk verification sample set. These samples cover the first and second risk characteristics and contain the labels of the related financing guarantee risks. This verification sample set will be used to evaluate the performance and generalization ability of the model on new data. The training sample set of the first risk characteristics, the training sample set of the second risk characteristics and the financing guarantee risk verification sample set are combined into a fusion sample set, which is used to train and verify the fusion model.
[0062] Use the fusion sample set as the training set, and adopt common machine learning training algorithms, such as gradient descent method, random forest, etc., for each initial sub-fusion model, and use the fusion sample set for training. After the training is completed, the m sub-fusion models are integrated. This can adopt simple voting, weighted voting, averaging and other integration methods to comprehensively utilize the prediction results of all models, and use the integrated fusion model as the final output. This integrated model integrates the prediction capabilities of all initial sub-fusion models and is used for financing guarantee risk assessment tasks.
[0063] Furthermore, the m initial sub-fusion models are trained according to the fusion sample set to output an integrated fusion model, and the method further includes:
[0064] The m initial sub-fusion models are trained according to the fusion sample set to obtain m financing guarantee risk output sample sets;
[0065] Based on the m financing guarantee risk output sample sets and the financing guarantee risk verification sample sets, an optimized sub-fusion model is obtained, wherein the probability accuracy of the optimized sub-fusion model is greater than or equal to a preset accuracy;
[0066] The weight ratio of the next round is adjusted according to the optimized sub-fusion model, and so on, to obtain m sub-fusion models after a preset number of iteration rounds;
[0067] The m sub-fusion models are integrated to obtain an integrated fusion model, and then a weight ratio λ1:λ2 of the integrated fusion model is output.
[0068] Take the fusion sample set as input, train each initial sub-fusion model, each model uses the features in the fusion sample set for training, and generates corresponding financing guarantee risk output sample sets according to its learning algorithm and parameters. These output sample sets reflect the risk assessment results of each sub-model on the samples in the fusion sample set. Repeat the above steps for all m initial sub-fusion models to obtain m financing guarantee risk output sample sets.
[0069] Using m financing guarantee risk output sample sets, by comparing with the financing guarantee risk verification sample sets, the probability accuracy and other performance indicators of each model are calculated, and a preset accuracy is set as the optimization target. This preset accuracy can be set according to specific needs and actual conditions. For the sub-fusion model whose probability accuracy is greater than or equal to the preset accuracy, it is extracted as the optimized sub-fusion model. This optimized model has better performance and generalization ability and can be used for subsequent financing guarantee risk assessment tasks.
[0070] According to the performance of the optimized sub-fusion model, the weight ratio of the next round is adjusted. According to the performance of each model, its weight in the ensemble can be adjusted to further improve the overall performance. Use the adjusted weight ratio to train the initial m sub-fusion models again with the fusion sample set. This step is similar to the previous training process, but the adjusted weight ratio is used here. Repeat the above steps until the preset iteration rounds are reached. In each iteration, the weight ratio is adjusted according to the optimized model performance, and the sub-fusion model is retrained. After completing the preset iteration rounds, the final m sub-fusion models are obtained. These models have better performance and generalization ability based on the weight adjustment and training optimization in the iteration process.
[0071] The m sub-fusion models are integrated by using simple voting, weighted voting, averaging and other integration methods. According to the performance of the integrated model, the weight ratio λ1:λ2 of the integrated fusion model is output, where λ1 represents the weight of the first type of risk feature, λ2 represents the weight of the second type of risk feature, and λ1<λ2.
[0072] Furthermore, the batch business data set is input into the business data transfer platform, and the method further includes:
[0073] Obtaining the unit byte quantity used for data transmission of each service data type in the batch service data set;
[0074] Performing real-time byte quantity analysis according to the unit byte quantity of each business data type, and setting a batch transmission instruction if the current real-time byte quantity is greater than a preset byte quantity, wherein the preset byte quantity is obtained by configuring parameters of the business data transfer platform;
[0075] The batch business data sets are input into the business data transfer platform in batches according to the batch transmission instruction.
[0076] Identify each business data type in the batch business data set, and for each business data type, determine its unit byte amount during the data transmission process. The unit byte amount represents the number of bytes occupied by each data unit. For example, for text data, the unit byte amount can be the number of bytes occupied by each character; for image data, it can be the number of bytes occupied by each pixel.
[0077] By real-time monitoring of data transmission flow, the current amount of business data being transmitted is monitored, and the total number of bytes transmitted, i.e., the real-time byte amount, is calculated in real time according to the unit byte amount of each business data type. The preset byte amount is obtained from the configuration parameters of the business data transfer platform. The preset byte amount is a pre-set threshold value, which indicates the maximum number of bytes allowed for transmission. When the real-time byte amount exceeds the preset byte amount, batch transmission is required. Once the real-time byte amount is greater than the preset byte amount, i.e., exceeds the set threshold value, the batch transmission instruction is triggered, instructing the platform to process and transmit the data in batches.
[0078] According to the batch transmission instructions, the batch business data set is processed in batches, and the data set is split into multiple batches according to the preset batch size. The batch-processed business data set is input into the business data transfer platform batch by batch until all batches of data are transmitted.
[0079] In summary, the method for dynamic monitoring and early warning of financing guarantee risks based on data fusion provided in the embodiment of the present application has the following technical effects:
[0080] 1. The business data transfer platform can be used to integrate business data sets from different sources, and use the pre-trained deadline-failure network model to perform data failure analysis, thereby achieving comprehensive integration and analysis of multi-source data and improving the accuracy and comprehensiveness of monitoring results;
[0081] 2. By extracting features from the first batch business data set and the second batch business data set, the risk features associated with the financing guarantee risk quantification are obtained, and risk assessment is performed in combination with the embedded financing guarantee risk assessment model, which can fully consider the impact of multiple risk factors and improve the accuracy and reliability of the assessment results;
[0082] 3. Use the embedded financing guarantee risk assessment model to conduct financing guarantee risk assessment, and make judgments through the set preset financing guarantee risk indicators. When the risk indicator exceeds the preset value, it can generate early warning signals in time, thereby improving the timeliness and accuracy of monitoring and early warning.
[0083] In summary, this dynamic monitoring and early warning method for financing guarantee risks based on data fusion improves the accuracy, comprehensiveness and timeliness of risk monitoring and early warning by integrating multi-source data, comprehensively analyzing risk factors, and combining real-time monitoring and early warning mechanisms.
[0084] Based on the same inventive concept as the method for dynamic monitoring and early warning of financing guarantee risks based on data fusion in the aforementioned embodiment, Figure 2 As shown, the present application provides a dynamic monitoring and early warning system for financing guarantee risks based on data fusion, the system comprising:
[0085] A transfer platform building module 10, wherein the transfer platform building module 10 is used to build a business data transfer platform, wherein the business data transfer platform is a transfer platform for business data transmission between the first server and the second server;
[0086] A business data input module 20, the business data input module 20 is used to monitor batch business data sets associated with financing guarantee risk quantification from the first service end, and input the batch business data sets into the business data transfer platform;
[0087] A business data identification module 30, wherein the business data identification module 30 is used by the business data transfer platform to identify the batch business data set and determine the first type of batch business data set and the second type of batch business data set, wherein the business data transfer platform includes a pre-trained deadline-failure network model, and the deadline-failure network model is used to perform data failure analysis on the batch business data set;
[0088] A feature extraction module 40, wherein the feature extraction module 40 is used to extract features from the first type of batch business data set and the second type of batch business data set respectively, to obtain first type of risk features and second type of risk features quantitatively associated with financing guarantee risks;
[0089] A risk indicator output module 50, wherein the risk indicator output module 50 is used for the business data transfer platform to transmit the first and second risk characteristics to the second server, and the second server uses the embedded financing guarantee risk assessment model to output the financing guarantee risk indicator;
[0090] The warning signal generating module 60 is used to generate a first warning signal when the financing guarantee risk index is greater than a preset financing guarantee risk index.
[0091] Furthermore, the system further includes a batch business data set identification module to perform the following operation steps:
[0092] Decomposing the business data types of the batch business data set to obtain each business data type;
[0093] Perform failure index analysis on each service data type according to the deadline-failure network model, and output failure factors of each service data type;
[0094] The batch business data sets are classified using the failure factors of various business data types as variables to determine a first type of batch business data sets whose failure factors are greater than or equal to a preset failure factor and a second type of batch business data sets whose failure factors are less than the preset failure factor.
[0095] Furthermore, the system further includes a network model output module to perform the following operation steps:
[0096] Get the fully connected neural network layer;
[0097] Obtaining a training sample data set, wherein the training sample data set includes batch business sample data, multiple failure penalty factors, and identification information for identifying the failure factors, wherein the multiple failure penalty factors include deadline-data credibility, deadline-data accuracy, and deadline-data value;
[0098] The fully connected neural network layer is trained according to the training sample data set until the model is in a convergence state, and the deadline-failure network model is output.
[0099] Furthermore, the second service end uses the embedded financing guarantee risk assessment model to perform risk assessment on the first risk characteristics and the second risk characteristics, and outputs a financing guarantee risk index;
[0100] Among them, the first type of risk characteristics and the second type of risk characteristics include an optimal weight ratio, and the weight of the first type of risk characteristics is less than the weight of the second type of risk characteristics, and risk assessment is performed based on the optimal weight ratio in combination with the first type of risk characteristics and the second type of risk characteristics.
[0101] Furthermore, the system further includes an integrated fusion model output module to perform the following operation steps:
[0102] Obtain m initial sub-fusion models, and assign m different weight ratios λ1:λ2 to the m initial sub-fusion models, wherein the weight λ1 of the first type of risk feature is less than the weight λ2 of the second type of risk feature;
[0103] Constructing a fusion sample set, the fusion sample set comprising a training sample set of the first type of risk characteristics, a training sample set of the second type of risk characteristics, and a financing guarantee risk verification sample set of the first type of risk characteristics and the second type of risk characteristics;
[0104] The m initial sub-fusion models are trained according to the fusion sample set, and an integrated fusion model is output.
[0105] Furthermore, the system further includes a sub-fusion model integration module to perform the following operation steps:
[0106] The m initial sub-fusion models are trained according to the fusion sample set to obtain m financing guarantee risk output sample sets;
[0107] Based on the m financing guarantee risk output sample sets and the financing guarantee risk verification sample sets, an optimized sub-fusion model is obtained, wherein the probability accuracy of the optimized sub-fusion model is greater than or equal to a preset accuracy;
[0108] The weight ratio of the next round is adjusted according to the optimized sub-fusion model, and so on, to obtain m sub-fusion models after a preset number of iteration rounds;
[0109] The m sub-fusion models are integrated to obtain an integrated fusion model, and then a weight ratio λ1:λ2 of the integrated fusion model is output.
[0110] Furthermore, the system further includes a batch business data set input module to perform the following operation steps:
[0111] Obtaining the unit byte quantity used for data transmission of each service data type in the batch service data set;
[0112] Performing real-time byte quantity analysis according to the unit byte quantity of each business data type, and setting a batch transmission instruction if the current real-time byte quantity is greater than a preset byte quantity, wherein the preset byte quantity is obtained by configuring parameters of the business data transfer platform;
[0113] The batch business data sets are input into the business data transfer platform in batches according to the batch transmission instruction.
[0114] Through the aforementioned detailed description of the method for dynamic monitoring and early warning of financing guarantee risks based on data fusion in this specification, those skilled in the art can clearly understand the dynamic monitoring and early warning system for financing guarantee risks based on data fusion in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.
[0115] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in FIG. Figure 3 As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store news data and data such as time decay factors. The network interface of the computer device is used to communicate with an external terminal through a network connection. The computer program is executed by the processor to implement a dynamic monitoring and early warning method for financing guarantee risks based on data fusion.
[0116] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0117] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the following steps when executing the computer program: building a business data transfer platform, wherein the business data transfer platform is a transfer platform for business data transmission between a first server and a second server; monitoring batch business data sets associated with financing guarantee risk quantification from the first server, and inputting the batch business data sets into the business data transfer platform; the business data transfer platform identifies the batch business data sets, determines a first type of batch business data set and a second type of batch business data set, wherein the business data transfer platform includes a pre-trained term-failure network model, and the term-failure network model is used to perform data failure analysis on the batch business data sets; feature extraction is performed on the first type of batch business data set and the second type of batch business data set, respectively, to obtain a first type of risk feature and a second type of risk feature associated with financing guarantee risk quantification; the business data transfer platform transmits the first type of risk feature and the second type of risk feature to the second server, and the second server outputs a financing guarantee risk index using an embedded financing guarantee risk assessment model; when the financing guarantee risk index is greater than a preset financing guarantee risk index, a first warning signal is generated.
[0118] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: building a business data transfer platform, wherein the business data transfer platform is a transfer platform for transmitting business data between a first server and a second server; monitoring batch business data sets associated with financing guarantee risk quantification from the first server, and inputting the batch business data sets into the business data transfer platform; the business data transfer platform identifies the batch business data sets, determines a first type of batch business data set and a second type of batch business data set, wherein the business data transfer platform includes a pre-trained term-failure network model, and the term-failure network model is used to perform data failure analysis on the batch business data sets; feature extraction is performed on the first type of batch business data set and the second type of batch business data set, respectively, to obtain a first type of risk feature and a second type of risk feature associated with financing guarantee risk quantification; the business data transfer platform transmits the first type of risk feature and the second type of risk feature to the second server, and the second server uses an embedded financing guarantee risk assessment model to output a financing guarantee risk index; when the financing guarantee risk index is greater than a preset financing guarantee risk index, a first warning signal is generated.
[0119] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0120] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. 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 application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A dynamic monitoring and early warning method for financing guarantee risks based on data fusion, characterized in that: The method comprises: Building a business data transfer platform, wherein the business data transfer platform is a transfer platform for business data transmission between the first server and the second server; Monitoring a batch of business data sets associated with the quantification of financing guarantee risks from the first service end, and inputting the batch of business data sets into the business data transfer platform; The business data transfer platform identifies the batch business data set to determine a first type of batch business data set and a second type of batch business data set, wherein the business data transfer platform includes a pre-trained deadline-failure network model, and the deadline-failure network model is used to perform data failure analysis on the batch business data set; Extracting features from the first type of batch business data set and the second type of batch business data set respectively to obtain first type of risk features and second type of risk features quantitatively associated with financing guarantee risks; The business data transfer platform transmits the first and second risk characteristics to the second server, and the second server outputs the financing guarantee risk index by using the embedded financing guarantee risk assessment model; When the financing guarantee risk indicator is greater than a preset financing guarantee risk indicator, a first warning signal is generated.
2. The method according to claim 1, characterized in that The business data transfer platform identifies the batch business data set to determine a first type of batch business data set and a second type of batch business data set, the method comprising: Decomposing the business data type of the batch business data set to obtain each business data type; Perform failure index analysis on each service data type according to the deadline-failure network model, and output failure factors of each service data type; The batch business data sets are classified using the failure factors of various business data types as variables to determine a first type of batch business data sets whose failure factors are greater than or equal to a preset failure factor and a second type of batch business data sets whose failure factors are less than the preset failure factor.
3. The method according to claim 2, characterized in that The method further comprises: Get the fully connected neural network layer; Obtaining a training sample data set, wherein the training sample data set includes batch business sample data, multiple failure penalty factors, and identification information for identifying the failure factors, wherein the multiple failure penalty factors include deadline-data credibility, deadline-data accuracy, and deadline-data value; The fully connected neural network layer is trained according to the training sample data set until the model is in a convergence state, and the deadline-failure network model is output.
4. The method according to claim 1, characterized in that The second server uses the embedded financing guarantee risk assessment model to perform risk assessment on the first and second risk characteristics, and outputs a financing guarantee risk index; Among them, the first type of risk characteristics and the second type of risk characteristics include an optimal weight ratio, and the weight of the first type of risk characteristics is less than the weight of the second type of risk characteristics, and risk assessment is performed based on the optimal weight ratio in combination with the first type of risk characteristics and the second type of risk characteristics.
5. The method according to claim 4, characterized in that The method for obtaining the optimization weight ratio includes: Obtain m initial sub-fusion models, and assign m different weight ratios λ1:λ2 to the m initial sub-fusion models, wherein the weight λ1 of the first type of risk feature is less than the weight λ2 of the second type of risk feature; Constructing a fusion sample set, the fusion sample set comprising a training sample set of the first type of risk characteristics, a training sample set of the second type of risk characteristics, and a financing guarantee risk verification sample set of the first type of risk characteristics and the second type of risk characteristics; The m initial sub-fusion models are trained according to the fusion sample set, and an integrated fusion model is output.
6. The method according to claim 5, characterized in that The m initial sub-fusion models are trained according to the fusion sample set to output an integrated fusion model, and the method further includes: The m initial sub-fusion models are trained according to the fusion sample set to obtain m financing guarantee risk output sample sets; Based on the m financing guarantee risk output sample sets and the financing guarantee risk verification sample sets, an optimized sub-fusion model is obtained, wherein the probability accuracy of the optimized sub-fusion model is greater than or equal to a preset accuracy; The weight ratio of the next round is adjusted according to the optimized sub-fusion model, and so on, to obtain m sub-fusion models after a preset number of iteration rounds; The m sub-fusion models are integrated to obtain an integrated fusion model, and then a weight ratio λ1:λ2 of the integrated fusion model is output.
7. The method according to claim 1, characterized in that Inputting the batch business data set into the business data transfer platform, the method further includes: Obtaining the unit byte quantity used for data transmission of each service data type in the batch service data set; Performing real-time byte quantity analysis according to the unit byte quantity of each business data type, and setting a batch transmission instruction if the current real-time byte quantity is greater than a preset byte quantity, wherein the preset byte quantity is obtained by configuring parameters of the business data transfer platform; The batch business data sets are input into the business data transfer platform in batches according to the batch transmission instruction.
8. The dynamic monitoring and early warning system for financing guarantee risks based on data fusion is characterized by: The system for implementing the method for dynamic monitoring and early warning of financing guarantee risks based on data fusion according to any one of claims 1 to 7 comprises: A transfer platform building module, wherein the transfer platform building module is used to build a business data transfer platform, wherein the business data transfer platform is a transfer platform for business data transmission between the first server and the second server; A business data input module, the business data input module is used to monitor batch business data sets associated with financing guarantee risk quantification from the first service end, and input the batch business data sets into the business data transfer platform; A business data identification module, wherein the business data identification module is used by the business data transfer platform to identify the batch business data set and determine the first type of batch business data set and the second type of batch business data set, wherein the business data transfer platform includes a pre-trained deadline-failure network model, and the deadline-failure network model is used to perform data failure analysis on the batch business data set; A feature extraction module, the feature extraction module is used to extract features from the first type of batch business data set and the second type of batch business data set respectively, to obtain first type of risk features and second type of risk features quantitatively associated with financing guarantee risks; A risk indicator output module, wherein the risk indicator output module is used for the business data transfer platform to transmit the first and second risk characteristics to the second server, and the second server uses the embedded financing guarantee risk assessment model to output the financing guarantee risk indicator; The early warning signal generation module is used to generate a first early warning signal when the financing guarantee risk index is greater than a preset financing guarantee risk index.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method for dynamic monitoring and early warning of financing guarantee risks based on data fusion described in any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the steps of the method for dynamic monitoring and early warning of financing guarantee risks based on data fusion described in any one of claims 1 to 7 are implemented.
Citation Information
Patent Citations
Asset allocation risk assessment method and device, computer equipment and storage medium
CN110135695A
Meta-learning method and device, initialization method, computing equipment and storage medium
CN110852447A
Risk control system and method for financing guarantee industry
CN116385169A
Risk control model training method, risk prediction method and related device
CN116703558A
Enterprise violation data detection and identification method fusing knowledge graph and graph attention network
CN117494048A
