Automatic risk management method and device, equipment and medium

Abstract

The invention relates to the technical field of medical science and technology and risk prediction, and discloses an automatic risk management method, device and equipment and a medium, and the method comprises the steps: receiving historical data, analyzing the historical data, extracting a basic feature set and a label from the historical data, connecting the basic feature set with the label, and building a basic feature library of a system; presetting a risk scene, selecting a specific feature set from the basic feature library according to the risk scene, and automatically constructing a training sample set according to the specific feature set; and training the training sample set based on machine learning and constructing a risk prediction model. Full-process automatic management of historical case backtracking, current risk detection and management and future risk prediction is integrated, separation among functional modules of a traditional system is broken through, the overall risk management efficiency is improved, unsupervised, semi-supervised and full-supervised learning modes are supported, modeling requirements in different risk scenes are met, and the system is suitable for large-scale popularization and application. A proper algorithm can be automatically selected for model training according to a service scene.

Description

Technical Field

[0001] The present invention relates to the fields of medical technology and risk prediction technology, and in particular to an automated risk management method, device, equipment and medium. Background Art

[0002] The existing risk management and control systems in the insurance industry have single functions and cannot fully cover the detection and tracing of existing risks, the detection and prevention of current risks, and the prediction of future risks. Different types of risk management systems are often fragmented and cannot be effectively integrated. In addition, the existing systems are inefficient in modeling emerging risks, lack automation and rapid response mechanisms, and perform poorly in dealing with unknown risks. This makes it impossible for insurance companies to respond to the increasingly complex risk environment in a timely and comprehensive manner. Summary of the invention

[0003] The present invention provides an automated risk management method, apparatus, computer equipment and medium to solve the technical problem that the risk management and control system has a single function and is difficult to fully cover the detection and backtracking of risks that have occurred, the detection and prevention of current risks, and the prediction of future risks.

[0004] In a first aspect, an automated risk management method is provided, comprising:

[0005] Receive historical data, analyze historical data, extract basic feature sets and labels from historical data, and establish the basic feature library of the system with basic feature sets and labels;

[0006] Preset risk scenarios, select specific feature sets from the basic feature library according to the risk scenarios, and automatically construct training sample sets based on the specific feature sets;

[0007] Train the sample set based on machine learning and build a risk prediction model.

[0008] In a second aspect, an automated risk management device is provided, comprising:

[0009] The receiving module is used to receive historical data, analyze historical data, extract basic feature sets and labels from historical data, and establish a basic feature library of the system with the basic feature sets and labels;

[0010] The sample construction module is used to preset risk scenarios, select specific feature sets from the basic feature library according to the risk scenarios, and automatically construct training sample sets according to the specific feature sets;

[0011] The training module is used to train the training sample set based on machine learning and build a risk prediction model.

[0012] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned automated risk management method when executing the computer program.

[0013] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned automated risk management method are implemented.

[0014] In the scheme implemented by the above-mentioned automated risk management method, device, computer equipment and storage medium, it is possible to receive historical data, analyze historical data, extract basic feature sets and labels from historical data, and establish the basic feature library of the system by combining the basic feature sets and labels; preset risk scenarios, select specific feature sets from the basic feature library according to the risk scenarios, and automatically construct training sample sets according to the specific feature sets; train the training sample sets based on machine learning and build risk prediction models. It integrates the full-process automated management of historical case backtracking, current risk detection and management, and future risk prediction, breaking the separation between the functional modules of the traditional system, improving the overall risk management efficiency, supporting unsupervised, semi-supervised and fully supervised learning modes, adapting to the modeling needs under different risk scenarios, and being able to automatically select appropriate algorithms for model training according to business scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0016] Figure 1 is a schematic diagram of an application environment of an automated risk management method in one embodiment of the present invention;

[0017] Figure 2 is a flow chart of an automated risk management method in one embodiment of the present invention;

[0018] Figure 3 is a schematic diagram of a structure of an automated risk management device in one embodiment of the present invention;

[0019] Figure 4 is a schematic diagram of a structure of a computer device in one embodiment of the present invention;

[0020] Figure 5 It is another structural schematic diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The automated risk management method provided by the embodiment of the present invention can be applied in Figure 1 In an application environment, the client communicates with the server through a network. The server can receive historical data through the client, analyze the historical data, extract basic feature sets and labels from the historical data, and establish a basic feature library of the system with the basic feature sets and labels; preset risk scenarios, select specific feature sets from the basic feature library according to the risk scenarios, and automatically construct training sample sets according to the specific feature sets; train the training sample sets based on machine learning and build a risk prediction model. In the present invention, for complex insurance entities such as medical insurance business, the full-process automated management of historical case backtracking, current risk detection and management, and future risk prediction is integrated, breaking the separation between the functional modules of the traditional system, improving the overall risk management efficiency, supporting unsupervised, semi-supervised and fully supervised learning modes, adapting to the modeling requirements under different risk scenarios, and being able to automatically select appropriate algorithms for model training according to business scenarios.

[0023] See also Figure 2 As shown, Figure 2 A flow chart of an automated risk management method provided by an embodiment of the present invention includes the following steps:

[0024] S10: receiving historical data, analyzing the historical data, extracting basic feature sets and labels from the historical data, and establishing a basic feature library of the system with the basic feature sets and labels;

[0025] The automated risk management method provided by the present invention can be applied to the risk management of the medical insurance industry in various application scenarios. It can realize the retrospection of historical cases, the detection and early warning of current risks, and the prediction of future risks based on business needs, dynamically adjust and optimize the risk model, realize the automated management of the whole process, optimize the business process, and improve the efficiency of risk management. The automated risk management method is applicable to both financial and medical scenarios.

[0026] The basic feature library of the system is generated through the feature set of historical data. Business experts can build customized features based on business needs to form a feature set. The update and expansion of the basic feature library provides a solid data foundation for subsequent risk modeling.

[0027] The system analyzes existing business data (such as underwriting data, claims data, customer data, etc.) and extracts basic feature sets and labels;

[0028] Business experts define custom features based on business needs through the system's interactive interface to form a feature set. Custom features will be dynamically adjusted as data is continuously updated to better meet the needs of real-time business scenarios.

[0029] The system supports online updating of the feature library. That is, as business data grows, the system automatically extracts valid information from the new data and adds it to the basic feature library to ensure the accuracy and timeliness of the model.

[0030] Receive historical data, analyze historical data specifically: clarify the needs and goals of feature values, so as to determine which historical data to collect. Historical data can come from a variety of channels, such as databases, log files, questionnaires, public data sources, etc. When collecting data, ensure the integrity, accuracy and timeliness of the data. Of course, before that, data preprocessing is also required: data cleaning: remove duplicate data, handle missing values, correct erroneous data, etc., convert data into a format suitable for analysis, such as converting text data into numerical data, or standardizing and normalizing data. Data integration: merge data from different sources to ensure data consistency and integrity.

[0031] Then, through data analysis and data mining techniques, such as principal component analysis (PCA), cluster analysis, association rule mining, etc., a representative and discriminative basic feature set is extracted from the historical data. The feature extraction process needs to be continuously iterated and optimized to ensure that the extracted basic feature set can effectively reflect the essential attributes of the historical data.

[0032] The extracted basic feature sets and labels are sorted and organized to form a feature library of the system. The basic feature library can include information such as the name, type, description, value range, importance, etc. of the feature. The basic feature library should be extensible and maintainable to facilitate the subsequent addition of new features and updating of existing features.

[0033] In the basic feature library, not all features are useful for the task. Therefore, feature selection is required to select the most valuable features for the task. Feature selection methods include filtering, wrapping, and embedding. Through feature selection, the complexity of the model can be reduced and the performance and interpretability of the model can be improved. At the same time, it is also necessary to optimize the features, such as feature scaling and feature encoding, to improve the quality of the features and the accuracy of the model.

[0034] In order to facilitate the management of the basic feature library, a management mechanism for the basic feature library is established, including feature version control, permission management, data backup, etc. The basic feature library is applied to subsequent data analysis and machine learning tasks to improve the efficiency and accuracy of the tasks. As time goes by and data changes, the basic feature library is continuously updated and optimized to adapt to new business needs and data characteristics.

[0035] S20: Preset a risk scenario, select a specific feature set from the basic feature library according to the risk scenario, and automatically construct a training sample set according to the specific feature set;

[0036] Business experts can select specific feature sets based on risk scenarios, and the system will automatically construct training sample sets and conduct model training. For unsupervised scenarios, the system does not require sample labels; for scenarios with a small number of samples, the system automatically performs semi-supervised learning to complete model training. Specifically, business experts can select specific risk scenarios through the system, such as emerging insurance fraud risks or claims risks of collusion between specific gangs. The system will automatically call the corresponding specific feature sets to generate training sample sets;

[0037] It should be noted that, on the basis of clarifying the risk categories (such as credit risk, market risk, operational risk, etc.), the risk scenarios are further refined, including the triggering conditions, potential impacts, historical cases, etc. of risk events. For example, for credit risk, it can be refined to the specific risk scenario of "individual consumers defaulting on loans due to income fluctuations". The key elements in the risk scenario are matched with the features in the basic feature library to identify the features that are highly relevant to risk prediction. Through feature selection algorithms (such as model-based feature importance scoring, mutual information method, etc.), the specific feature set that has the greatest impact on risk prediction is selected from a large number of features.

[0038] Review and sort out the basic feature library to ensure that the feature library covers multiple dimensions such as basic customer information, transaction behavior, credit record, external environment, etc. The basic feature library regularly introduces new features and removes outdated or redundant features. Further optimize the specific feature set that has been initially screened, remove redundant features, reduce collinearity between features, and improve the stability and predictive ability of the specific feature set. After that, further streamline the feature set through feature dimension reduction techniques (such as principal component analysis PCA, linear discriminant analysis LDA, etc.). Verify the optimized specific feature set with historical data to evaluate the accuracy, stability and robustness of the specific feature set in predicting risk events. If the verification result is not ideal, the feature selection strategy needs to be readjusted until a satisfactory specific feature set is obtained. Based on the risk case records in the historical data, annotate the specific feature set, generate a labeled training data set, and construct a training sample set using the annotated data and the optimized feature set.

[0039] S30: Train the sample set based on machine learning and build a risk prediction model.

[0040] Extract features that have an important impact on risk prediction from the original data and screen them through correlation analysis, mutual information, recursive feature elimination and other methods. Construct new features (such as lag features and combined features of time series data) according to business logic and data characteristics to enhance the expressiveness of the model. Scaling processing is performed on numerical features, such as standardization (making the features have unit variance) or normalization (making the feature values ​​fall within the interval [0,1]) to eliminate the influence of different feature dimensions. According to the nature of the problem and the characteristics of the data, select appropriate machine learning algorithms, such as logistic regression, decision tree, random forest, support vector machine, neural network, etc. For complex risk prediction problems, deep learning models (such as convolutional neural networks, recurrent neural networks, Transformer, etc.) are more effective. Use the training set data to train the selected model and optimize the performance of the model by adjusting the model's hyperparameters (such as learning rate, number of iterations, regularization parameters, etc.). Cross-validation and other techniques can be used during training to reduce the risk of overfitting.

[0041] Based on the training sample set of semi-supervised learning in machine learning, a risk prediction model is constructed according to the training results. The system supports the selection of multiple machine learning models, including unsupervised learning, semi-supervised learning and supervised learning models. Unsupervised models can discover potential risk patterns in data without clear labels, while semi-supervised learning models can use a small number of labeled samples for training to improve the generalization ability of the model. The system supports unsupervised, semi-supervised and fully supervised learning modes, adapts to the modeling needs under different risk scenarios, and can automatically select appropriate algorithms for model training according to business scenarios.

[0042] Semi-supervised learning uses the distribution information of unlabeled data to assist learning, thereby improving the performance of the model. Specifically, for labeled data, the algorithm uses traditional supervised learning methods to update the model weights; for unlabeled data, the algorithm learns by minimizing the prediction differences between other similar training examples, or utilizes these unlabeled data through other means (such as self-training, transfer learning, etc.).

[0043] Unsupervised learning uses unlabeled data to train models and discover hidden patterns and structures from the data without relying on predefined labels or results. Unlike supervised learning, the training samples of unsupervised learning do not have clear labels or known outcome variables. The model needs to learn from the data and discover its inherent structure and patterns without any explicit guidance or labels. Unsupervised learning focuses more on the understanding and analysis of the data itself, and its core goal is to explore the inherent structure and laws of the data. Since no specific target variables are required, unsupervised learning can be more flexibly applied to a variety of problems. It can not only be used for tasks such as data clustering and dimensionality reduction, but also for scenarios such as anomaly detection and data generation.

[0044] During the training process, the system will automatically select the most suitable algorithm based on the input business scenario, and optimize the model in the background to form the final risk prediction model. After the model training is completed, the system will apply it to current and future data to generate risk warnings and related decision-making recommendations.

[0045] Retrospective cases are collected by collecting abnormal cases and establishing an abnormal case feature library; an advanced feature library is formed by combining the abnormal case feature library and the basic feature library; when a preset risk scenario is obtained, historical cases matching the preset risk scenario are queried in the advanced feature library; detailed information of historical cases is retrospected and a retrospective report is generated.

[0046] The system supports the construction of features of abnormal cases, automatically searches for similar risk cases in historical data, and performs risk backtracking to help business personnel quickly locate potential risks. Through intelligent algorithms, the system can also identify potential risk patterns, generate early warnings in a timely manner, and respond to risks in advance.

[0047] Specifically, the system uses the anomaly detection module to discover abnormal cases in current business, automatically generate features of these cases, and search for similar risk cases in historical data. This function helps business experts trace back past risk cases, analyze the causes of the cases, and prevent similar current cases. The system supports in-depth analysis of historical cases, not only tracing back the process of the case, but also generating retrospective reports based on risk amount, case type and other dimensions, providing a basis for future risk management.

[0048] Case label classification, labeling risk types, where risk types include high-risk cases, low-risk cases and no-risk cases; predict the risk type of the target case based on the risk prediction model, and label the risk type of the target case.

[0049] The system has a powerful risk visualization function. Business experts can intuitively view the detailed information of risk cases on the interactive interface, including the distribution of similar cases, the timeline of case occurrence, etc. When business experts handle risk cases, they can label the system's prediction results and mark which cases are high-risk, low-risk or no-risk cases. The system will optimize the model in real time based on these labels, making future risk predictions more accurate;

[0050] When backtracking, if there are no risk cases to backtrack in the model, then by obtaining new risk cases, query similar cases in the risk prediction model; extract the case information of similar cases, and use the case information of similar cases as the case information of new risk cases, and update the case information of new risk cases in the risk prediction model. Experts can directly report newly discovered risk cases through the system, and the system will automatically search for similar cases and update the risk feature library. This function helps to deal with new risk types and improve the response speed to new risks.

[0051] Visualize risk prediction results, preset typical risk scenarios, and extract typical risk case information, risk factors and risk prediction results corresponding to typical risk scenarios from the risk prediction model; send typical risk case information, risk factors and risk prediction results to the visualization interface for display. In the visualization interface, staff can select specific risk cases for detailed analysis, and the system will display relevant risk factors, model prediction results, etc. to help experts make decisions; the system supports multi-department collaboration, and staff can report marked high-risk cases to relevant departments, and relevant departments can take risk management measures accordingly, such as risk amount recovery, case investigation, etc. The system provides a visual display function for risk cases, including backtracking of similar cases, future risk warnings, etc., to help business experts intuitively analyze risks and make quick decisions.

[0052] The system also provides a risk map that shows the distribution of different risks in geographical areas or time dimensions, allowing managers to see the overall trend of risks more intuitively. By setting up a risk map,

[0053] Among them, the risk map at least includes a geographical area; combining the risk prediction model with the risk map, the distribution of different risks in the geographical area is set.

[0054] It can be seen that in the above scheme, for complex insurance entities such as medical insurance business, first receive historical data, analyze historical data, extract basic feature sets and labels from historical data, and establish the basic feature library of the system by combining basic feature sets and labels; preset risk scenarios, select specific feature sets from the basic feature library according to risk scenarios, and automatically construct training sample sets according to specific feature sets; train training sample sets based on machine learning and build risk prediction models. It integrates the full process of automated management of historical case backtracking, current risk detection and management, and future risk prediction, breaking the separation between the functional modules of the traditional system, improving the overall risk management efficiency, supporting unsupervised, semi-supervised and fully supervised learning modes, adapting to the modeling needs under different risk scenarios, and being able to automatically select appropriate algorithms for model training according to business scenarios.

[0055] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0056] In one embodiment, an automated risk management device is provided, and the automated risk management device corresponds one-to-one to the automated risk management method in the above embodiment. Figure 3 As shown, the automated risk management device includes a receiving module 101, a sample construction module 102, and a training module 103. The detailed description of each functional module is as follows:

[0057] The receiving module 101 is used to receive historical data, analyze the historical data, extract basic feature sets and labels from the historical data, and establish a basic feature library of the system with the basic feature sets and labels;

[0058] The sample construction module 102 is used to preset risk scenarios, select specific feature sets from the basic feature library according to the risk scenarios, and automatically construct a training sample set according to the specific feature sets;

[0059] The training module 103 is used to train the training sample set based on machine learning and build a risk prediction model.

[0060] In one embodiment, the apparatus further comprises:

[0061] The collection module is used to collect abnormal cases and establish an abnormal case feature library;

[0062] The combination module is used to combine the abnormal case feature library and the basic feature library to form an advanced feature library;

[0063] The query matching case module is used to query the historical cases matching the preset risk scenario in the advanced feature library when the preset risk scenario is obtained;

[0064] The retrospective module is used to trace back the detailed information of historical cases and generate retrospective reports.

[0065] In one embodiment, the apparatus further comprises:

[0066] A labeling module is used to label risk types, where risk types include high-risk cases, low-risk cases, and no-risk cases;

[0067] The risk type prediction module is used to predict the risk type of the target case based on the risk prediction model and label the risk type of the target case.

[0068] In one embodiment, the apparatus further comprises:

[0069] The similar case query module is used to obtain new risk cases and query similar cases in the risk prediction model;

[0070] The extraction module is used to extract the case information of similar cases, and use the case information of similar cases as the case information of new risk cases, and update the case information of new risk cases in the risk prediction model.

[0071] In one embodiment, the apparatus further comprises:

[0072] The preset typical risk module is used to preset typical risk scenarios and extract typical risk case information, risk factors and risk prediction results corresponding to the typical risk scenarios from the risk prediction model;

[0073] The display module is used to send typical risk case information, risk factors and risk prediction results to the visualization interface for display.

[0074] In one embodiment, the apparatus further comprises:

[0075] A map setting module, used to set a risk map, wherein the risk map includes at least a geographical area;

[0076] The distribution module is used to combine the risk prediction model with the risk map to set the distribution of different risks in geographical areas.

[0077] The present invention provides an automated risk management device, which first receives historical data, analyzes historical data, extracts basic feature sets and labels from historical data, and establishes a basic feature library of the system by combining the basic feature sets and labels; presets risk scenarios, selects specific feature sets from the basic feature library according to the risk scenarios, and automatically constructs training sample sets according to the specific feature sets; trains the training sample sets based on machine learning and builds a risk prediction model. It integrates the full-process automated management of historical case backtracking, current risk detection and management, and future risk prediction, breaking the separation between the functional modules of the traditional system, improving the overall risk management efficiency, supporting unsupervised, semi-supervised and fully supervised learning modes, adapting to the modeling requirements under different risk scenarios, and being able to automatically select appropriate algorithms for model training according to business scenarios.

[0078] For the specific definition of the automated risk management device, please refer to the definition of the automated risk management method above, which will not be repeated here. Each module in the above-mentioned automated risk management device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0079] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or 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 network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, the functions or steps on the service side of an automated risk management method are implemented.

[0080] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device 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 and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server through a network connection. When the computer program is executed by the processor, it realizes the functions or steps of the client side of an automated risk management method

[0081] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0082] Receive historical data, analyze historical data, extract basic feature sets and labels from historical data, and establish the basic feature library of the system with basic feature sets and labels;

[0083] Preset risk scenarios, select specific feature sets from the basic feature library according to the risk scenarios, and automatically construct training sample sets based on the specific feature sets;

[0084] Train the sample set based on machine learning and build a risk prediction model.

[0085] 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:

[0086] Receive historical data, analyze historical data, extract basic feature sets and labels from historical data, and establish the basic feature library of the system with basic feature sets and labels;

[0087] Preset risk scenarios, select specific feature sets from the basic feature library according to the risk scenarios, and automatically construct training sample sets based on the specific feature sets;

[0088] Train the sample set based on machine learning and build a risk prediction model.

[0089] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0090] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0091] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0092] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An automated risk management method, characterized in that: include: Receive historical data, analyze historical data, extract basic feature sets and labels from historical data, and establish the basic feature library of the system with basic feature sets and labels; Preset risk scenarios, select specific feature sets from the basic feature library according to the risk scenarios, and automatically construct training sample sets based on the specific feature sets; Train the sample set based on machine learning and build a risk prediction model.

2. The method according to claim 1, characterized in that: The steps of training the sample set based on machine learning and building a risk prediction model include: The training sample set is trained based on semi-supervised learning in machine learning, and a risk prediction model is constructed according to the training results.

3. The method according to claim 2, characterized in that: After the step of training the training sample set based on machine learning and constructing a risk prediction model, the method further includes: Collect abnormal cases and establish abnormal case feature database; Combine the abnormal case feature library and the basic feature library to form an advanced feature library; When a preset risk scenario is obtained, the advanced feature library is searched for historical cases that match the preset risk scenario; Trace back the details of historical cases and generate traceback reports.

4. The method according to claim 1, characterized in that: After the step of training the training sample set based on machine learning and building a risk prediction model, the method further includes: Label risk types, where risk types include high-risk cases, low-risk cases, and no-risk cases; Predict the risk type of the target case based on the risk prediction model and label the risk type of the target case.

5. The method according to claim 1, characterized in that: After the step of training the training sample set based on machine learning and constructing a risk prediction model, the method further includes: Obtain new risk cases and query similar cases in the risk prediction model; The case information of similar cases is extracted, and the case information of similar cases is used as the case information of new risk cases, and the case information of new risk cases in the risk prediction model is updated.

6. The method according to claim 1, characterized in that: After the step of training the training sample set based on machine learning and constructing a risk prediction model, the method further includes: Preset typical risk scenarios and extract typical risk case information, risk factors and risk prediction results corresponding to the typical risk scenarios from the risk prediction model; Send typical risk case information, risk factors and risk prediction results to the visualization interface for display.

7. The method according to claim 1, characterized in that: After the step of training the training sample set based on machine learning and constructing a risk prediction model, the method further includes: Setting a risk map, wherein the risk map includes at least a geographical area; Combine the risk prediction model with the risk map to set the distribution of different risks in geographical areas.

8. An automated risk management device, characterized in that: include: The receiving module is used to receive historical data, analyze historical data, extract basic feature sets and labels from historical data, and establish a basic feature library of the system with the basic feature sets and labels; The sample construction module is used to preset risk scenarios, select specific feature sets from the basic feature library according to the risk scenarios, and automatically construct training sample sets according to the specific feature sets; The training module is used to train the training sample set based on machine learning and build a risk prediction model.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the automated risk management method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the automated risk management method according to any one of claims 1 to 7 are implemented.

Images