Method and device for determining business risk control model, equipment and storage medium

Abstract

The invention provides a business risk control model determination method and device, equipment and a storage medium. The method comprises the following steps: determining risk control strategy information of a target business and a risk control model set corresponding to a risk control strategy in the risk control strategy information; performing permutation and combination processing on the risk control models in the risk control model set to obtain a combined model set; determining a target combination model in the combination model set; and taking a target combination model in the combination model set as a risk control model in the risk control model set, re-executing the step of permutation and combination processing on the risk control models in the risk control model set to obtain the combination model set, and after the model combination number of the combination models in the obtained combination model set meets a preset number requirement, obtaining the risk control model set. Obtaining a global combination model set; and determining a target risk control combination model according to the global combination model set and the risk control strategy information. According to the method, the model accuracy of the business risk control model is improved, and the business risk control efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and particularly to a method, apparatus, device, and storage medium for determining a business risk control model. Background Art

[0002] With the rapid development of the economy, the security of risk control systems has received increasing attention. To ensure the safe operation of businesses, a business risk control system is usually adopted to detect the operation risks of businesses.

[0003] The existing risk control technologies mainly focus on the financial industry. By using big data means and relying on an efficient streaming computing engine, real-time processing of data is achieved to reach the ability of business risk control.

[0004] However, the existing technologies cannot achieve the efficient combination of business models and cannot achieve business risk control for industries other than the financial industry. Summary of the Invention

[0005] This application provides a method, apparatus, device, and storage medium for determining a business risk control model to solve the problem of risk control for business development and operation.

[0006] In a first aspect, this application provides a method for determining a business risk control model, including:

[0007] Determine the risk control strategy information of the target business and the set of risk control models corresponding to the risk control strategies in the risk control strategy information;

[0008] Perform permutation and combination processing on the risk control models in the set of risk control models to obtain a set of combined model groups. The set of combined model groups includes combined model groups, and each combined model group includes a combined model obtained by performing permutation and combination processing on the risk control models;

[0009] Determine the target combined model in the set of combined model groups. The target combined model is a combined model in the combined model group that meets the requirements of the risk control strategy;

[0010] Use the target combined model in the set of combined model groups as the risk control model in the set of risk control models, and re-execute the step of performing permutation and combination processing on the risk control models in the set of risk control models to obtain a set of combined model groups until the number of model combinations of the combined models in the obtained set of combined model groups meets the preset quantity requirement, and then obtain a global set of combined models;

[0011] Determine the target risk control combined model according to the global set of combined models and the risk control strategy information.

[0012] In the embodiments of this application, determining the risk control strategy information of the target business and the set of risk control models corresponding to the risk control strategies in the risk control strategy information includes:

[0013] Determine the target business and the business type of the target business;

[0014] According to the business type of the target business, determine the risk control strategy information of the target business;

[0015] According to the risk control strategy information, determine the risk control model set corresponding to the risk control strategy in the risk control strategy information.

[0016] In the embodiment of the present application, perform permutation and combination processing on the risk control models in the risk control model set to obtain a set of combined model groups, including:

[0017] According to the risk control strategy information of the target business, determine the preset number of models for permutation and combination;

[0018] According to the preset number of models, perform permutation and combination on the risk control models in the risk control model set to obtain each combined model group;

[0019] According to each combined model group, obtain the set of combined model groups.

[0020] In the embodiment of the present application, determine the target combined model in the set of combined model groups, including:

[0021] According to the set of combined model groups, determine the risk control effects of each combined model group in the set of combined model groups;

[0022] According to the risk control strategy information of the target business, determine the risk control strategy requirements in the risk control strategy information;

[0023] According to the risk control effects of each combined model group and the risk control strategy requirements, obtain the target combined model in the set of combined model groups that meets the risk control strategy requirements.

[0024] In the embodiment of the present application, after determining the risk control effects of each combined model group in the set of combined model groups according to the set of combined model groups, the method further includes:

[0025] According to the risk control effects of each combined model group, determine the risk control effects of each combined model in each combined model group;

[0026] Compare the risk control effects of each combined model with each other to obtain the initial target combined model in each combined model group;

[0027] According to the initial target combined model in each combined model group, determine the target combined model.

[0028] In the embodiment of the present application, after determining the target combined model in the set of combined model groups and before using the target combined model in the set of combined model groups as the risk control model in the risk control model set, the method further includes:

[0029] Determine the number of model combinations in the target combination model according to the target combination model;

[0030] Determine the preset quantity requirement of the model combination according to the number of risk control strategies in the risk control strategy information;

[0031] Compare the number of model combinations with the preset quantity requirement to obtain a quantity comparison result;

[0032] If the quantity comparison result is that the number of model combinations meets the preset quantity requirement, then obtain the global combination model set.

[0033] In the embodiment of the present application, after comparing the number of model combinations with the preset quantity requirement to obtain a quantity comparison result, the method further includes:

[0034] If the quantity comparison result is that the number of model combinations does not meet the preset quantity requirement, then use the target combination model in the combined model group set as the risk control model in the risk control model set, and re - execute the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain the combined model group set until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtain the global combination model set.

[0035] In the embodiment of the present application, determining the target risk control combination model according to the global combination model set and the risk control strategy information includes:

[0036] Determine each combined model in the global combination model set and the risk control effect of each combined model according to the global combination model set;

[0037] Compare the risk control effect of each combined model with the risk control strategy requirements in the risk control strategy information to determine the target risk control combination model that meets the risk control strategy requirements.

[0038] In a second aspect, the present application provides a device for determining a business risk control model, including:

[0039] A strategy information determination module, configured to determine the risk control strategy information of the target business and the risk control model set corresponding to the risk control strategy in the risk control strategy information;

[0040] A model group set obtaining module, configured to perform permutation and combination processing on the risk control models in the risk control model set to obtain a combined model group set, where the combined model group set includes combined model groups, and the combined model groups include combined models obtained by performing permutation and combination processing on the risk control models;

[0041] A combined model determination module, configured to determine the target combined model in the combined model group set, where the target combined model is a combined model in the combined model group that meets the risk control strategy requirements;

[0042] The combined model set obtaining module is used to take the target combined model in the combined model set as the risk control model in the risk control model set, and re - execute the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain a combined model group set, until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtain the global combined model set;

[0043] The target model determining module is used to determine the target risk control combined model according to the global combined model set and the risk control strategy information.

[0044] Thirdly, the present application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0045] The memory stores computer - executable instructions;

[0046] The processor executes the computer - executable instructions stored in the memory to implement the method of the embodiment of the present application.

[0047] Fourthly, the present application provides a computer - readable storage medium, in which computer - executable instructions are stored, and when the computer - executable instructions are executed by a processor, they are used to implement the method of the embodiment of the present application.

[0048] The method, device, equipment and storage medium for determining the business risk control model provided by the present application, by determining the risk control strategy information of the target business and the risk control model set corresponding to the risk control strategy in the risk control strategy information; performing permutation and combination processing on the risk control models in the risk control model set to obtain a combined model group set, the combined model group set includes combined model groups, and the combined model groups include the combined models obtained by performing permutation and combination processing on the risk control models; determining the target combined model in the combined model group set, the target combined model is the combined model in the combined model group that meets the risk control strategy requirements; taking the target combined model in the combined model group set as the risk control model in the risk control model set, and re - executing the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain a combined model group set, until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtaining the global combined model set; determining the target risk control combined model according to the global combined model set and the risk control strategy information, permuting and combining each model corresponding to the target business risk control strategy, so as to obtain a model combination with better risk control effect on the target business, realizing the effect of performing real - time risk control model combination, and avoiding the occurrence of risk problems brought during the handling process of the target business. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0050] Figure 1 It is a schematic flowchart of a method for determining a service risk control model provided by an embodiment of the present application;

[0051] Figure 2 It is a schematic flowchart of another method for determining a service risk control model provided by an embodiment of the present application;

[0052] Figure 3 It is a schematic structural diagram of a device for determining a service risk control model provided by an embodiment of the present application;

[0053] Figure 4 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application.

[0054] Through the above-mentioned drawings, specific embodiments of the present application have been shown, and there will be more detailed descriptions hereinafter. These drawings and text descriptions are not intended to limit the scope of the concept of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. Specific Embodiments

[0055] Here, exemplary embodiments will be described in detail, and examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0056] In the prior art, risk control products mainly focus on the risk control of financial capabilities, and the technical implementation mainly uses big data means, relying on efficient streaming computing engines such as spark and flink to achieve real-time data processing to achieve risk control capabilities. The logic is relatively fixed, and at the same time, the processing of voice and images cannot be supported, and AI intelligent processing cannot be achieved; at the same time, there is still a blank in fields with complex business logics such as operators, and existing financial risk control products cannot be applied to requirements such as manual authentication and semantic recognition, and there is also a lack of prediction of business development trends and risk control of the health of business development.

[0057] To solve the above problems, the method, device, equipment, and storage medium for determining a business risk control model provided by this application can determine corresponding risk control strategies and various models corresponding to the risk control strategies according to the risk control requirements of the target business. By performing permutations and combinations on each model, the target combined model with better risk control effect or the most compliant with the risk control strategy requirements among the combined models obtained after permutation and combination is determined. Then, according to the number of models in the target combined model, a global combined model set that meets the preset model number requirements in the model strategy requirements is determined from the target combined model, and a target risk control combined model with a risk control effect meeting the risk control strategy requirements is determined from this global combined model set. Thereby, the problem that the existing model risk control platform cannot achieve multi-model orchestration and integration according to the target business type is solved.

[0058] The following uses specific embodiments to elaborate in detail on the technical solution of this application and how the technical solution of this application solves the above technical problems. These several specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of this application will be described below with reference to the accompanying drawings.

[0059] The execution subject of the method for determining a business risk control model provided by an embodiment of this application can be a server. Among them, the server can be a device such as a computer. This embodiment does not make special restrictions on the implementation manner of the execution subject, as long as the execution subject can determine the risk control strategy information of the target business and the set of risk control models corresponding to the risk control strategies in the risk control strategy information; perform permutation and combination processing on the risk control models in the set of risk control models to obtain a set of combined model groups, the set of combined model groups includes combined model groups, and the combined model groups include the combined models obtained by performing permutation and combination processing on the risk control models; determine the target combined model in the set of combined model groups, and the target combined model is the combined model that complies with the risk control strategy requirements in the combined model group; use the target combined model in the set of combined model groups as the risk control model in the set of risk control models, and re-execute the step of performing permutation and combination processing on the risk control models in the set of risk control models to obtain a set of combined model groups until the number of model combinations of the combined models in the obtained set of combined model groups meets the preset number requirements, and then a global combined model set is obtained; determine the target risk control combined model according to the global combined model set and the risk control strategy information.

[0060] Among them, the target business refers to a business that needs to predict the business development trend and needs to control the business development, such as the operator business.

[0061] Figure 1 It is a schematic flowchart of the method for determining a business risk control model provided by an embodiment of this application. As Figure 1 shown, the method may include:

[0062] S101. Determine the risk control strategy information of the target business and the set of risk control models corresponding to the risk control strategies in the risk control strategy information.

[0063] Among them, the risk control strategy information refers to the strategy requirements needed to control the target business. For example, it is to judge the rationality of the business, whether the business is handled by the person himself / herself, and whether there is an issue of document abuse in the business.

[0064] The set of risk control models refers to each risk control model corresponding to the types of models required for risk control of the target business. For example, a face recognition model, a semantic analysis model, etc.

[0065] Among them, according to the business type of the target business, the business risk control system determines each risk control key point for risk control of the target business. According to each risk control key point, it determines each risk control model corresponding to each risk control key point and the number of models required to complete the overall risk control of the target business. For example, when the target business is the operator user account opening business, the business risk control system needs to determine risk control key points such as document abuse, whether it is handled by the person himself / herself, and business rationality required for business risk control according to this user account opening business. And according to the above risk control key points, it determines risk control models such as a face recognition model, a voice recognition model, a threshold prediction model, and an account opening business legality model required for risk control, and determines the specific model implementation method for completing this risk control. For example, the model implementation method can be to first determine the development rationality of the business, then determine whether it is handled by the person himself / herself, and then perform voice recognition and analysis on the person.

[0066] Among them, in the embodiment of this application, determining the risk control strategy information of the target business and the set of risk control models corresponding to the risk control strategies in the risk control strategy information includes:

[0067] Determine the target business and the business type of the target business;

[0068] According to the business type of the target business, determine the risk control strategy information of the target business;

[0069] According to the risk control strategy information, determine the set of risk control models corresponding to the risk control strategies in the risk control strategy information.

[0070] Among them, before determining the risk control combination model of the target business, the business risk control system has correspondingly configured each risk control strategy information of the target business according to the business type of the target business. At the same time, for each risk control model corresponding to each risk control strategy information, the corresponding model training and model configuration have been completed before determining the combination model to ensure that the confirmation of the target combination model of the target business can be completed.

[0071] S102. Perform permutation and combination processing on the risk control models in the risk control model set to obtain a combined model set. The combined model set includes combined model groups, and each combined model group includes a combined model obtained by performing permutation and combination processing on the risk control models.

[0072] Among them, the combined model set refers to a set of model combinations obtained by determining each model for completing the risk control of the target business according to the business risk control system and performing pairwise permutation and combination on each model.

[0073] For example, according to the business type of the target business, it is determined that the models that need to perform business risk control are the face recognition model, the voice recognition model, and the semantic understanding model; if the preset number of models for performing model permutation and combination in the risk control strategy information of the target business is two, then pairwise serial permutation and combination are performed on the above three models respectively. That is, pairwise serial permutation and combination are performed on the face recognition model and the voice recognition model, and a first combined model set between the face recognition model and the voice recognition model can be obtained. Among them, the first combined model set includes combined models of face recognition and face recognition, combined models of face recognition and voice recognition, combined models of voice recognition and voice recognition, and combined models of voice recognition and face recognition; pairwise serial combination is performed on the face recognition model and the semantic understanding model, and a second combined model set between the face recognition model and the semantic understanding model can be obtained, that is, combined models of face recognition and face recognition, combined models of face recognition and semantic understanding, combined models of semantic understanding and semantic understanding, and combined models of semantic understanding and face recognition; pairwise serial permutation and combination are performed on the voice recognition and semantic understanding models, and a third combined model set between the voice recognition and semantic understanding models can be obtained. Among them, the third combined model set includes combined models of voice recognition and voice recognition, combined models of voice recognition and semantic understanding, combined models of semantic understanding and semantic understanding, and combined models of semantic understanding and voice recognition; according to the above first combined model set, second combined model set, and third combined model set, the combined model set of the target business can be obtained.

[0074] Among them, in the embodiments of the present application, performing permutation and combination processing on the risk control models in the risk control model set to obtain a combined model set includes:

[0075] Determine the preset number of risk control models for permutation and combination according to the risk control strategy information of the target business;

[0076] Perform permutation and combination on the risk control models in the risk control model set according to the preset number of risk control models to obtain each combined model group;

[0077] Obtain the combined model set according to each combined model group.

[0078] Among them, the preset number of risk control models refers to the number of models required for each model permutation and combination preset in the risk control strategy information of the target business. For example, when the number of risk control models corresponding to the risk control strategy information is five, the preset number of risk control models can be greater than two and less than five. If the number of preset risk control models is two, then two models are required for pairwise permutation and combination during model permutation and combination to obtain the final combined model set.

[0079] S103. Determine the target combined model in the combined model set. The target combined model is the combined model in the combined model group that meets the requirements of the risk control strategy.

[0080] Among them, when the models required for target business risk control are face recognition model, speech recognition model, and semantic understanding model, the combined model set obtained after model permutation and combination includes the first model combination set between the face recognition model and the speech recognition model, the second model combination set between the face recognition model and the semantic understanding model, and the third model combination set between the speech recognition and semantic understanding models; comparing the risk control effects of each combined model in the first model combination set, the second model combination set, and the third model combination set with the requirements of the risk control strategy respectively, the first target model combination that meets the requirements of the risk control strategy in the first model combination set, the second target model combination that meets the requirements of the risk control strategy in the second model combination set, and the second target model combination that meets the requirements of the risk control strategy in the third model combination set can be obtained. The above first target model combination, second target model combination, and second target model combination are the target combined models in the combined model set.

[0081] Among them, in the embodiment of the present application, determining the target combined model in the combined model set includes:

[0082] According to the combined model set, determine the risk control effects of each combined model group in the combined model set;

[0083] According to the risk control strategy information of the target business, determine the risk control strategy requirements in the risk control strategy information;

[0084] According to the risk control effects of each combined model group and the risk control strategy requirements, obtain the target combined models in the combined model set that meet the requirements of the risk control strategy.

[0085] Among them, in the embodiment of the present application, after determining the risk control effects of each combined model group in the combined model set according to the combined model set, the method further includes:

[0086] According to the risk control effects of each combined model group, determine the risk control effects of each combined model in each combined model group;

[0087] Compare the risk control effects of each combined model with each other to obtain the initial target combined model in each combined model group;

[0088] Determine the target combined model according to the initial target combined model in each combined model group.

[0089] Among them, the initial target combined model refers to the combined model with relatively good risk control effect obtained after mutual comparison of each combined model in each combined model group.

[0090] Among them, when the models required for target business risk control are face recognition model, speech recognition model, and semantic understanding model, the set of combined model groups obtained after model permutation and combination includes the first model combination set between the face recognition model and the speech recognition model, the second model combination set between the face recognition model and the semantic understanding model, and the third model combination set between the speech recognition and semantic understanding models; comparing the risk control effects of each combined model in the first model combination set with each other, a first target combined model with relatively good risk control effect can be obtained; comparing the risk control effects of each combined model in the second model combination set with each other, a second target combined model with relatively good risk control effect can be obtained; comparing the risk control effects of each combined model in the third model combination set with each other, a third target combined model with relatively good risk control effect can be obtained; according to the first target combined model, the second target combined model, and the third target combined model, the final target combined model can be determined.

[0091] Among them, in the embodiment of the present application, after determining the target combined model in the combined model group set and before using the target combined model in the combined model group set as the risk control model in the risk control model set, the method further includes:

[0092] Determine the number of model combinations in the target combined model according to the target combined model;

[0093] Determine the preset quantity requirement of the model combination according to the number of risk control strategies in the risk control strategy information;

[0094] Compare the number of model combinations with the preset quantity requirement to obtain a quantity comparison result;

[0095] If the quantity comparison result is that the number of model combinations meets the preset quantity requirement, a global combined model set is obtained.

[0096] Among them, in the embodiment of the present application, after comparing the number of model combinations with the preset quantity requirement to obtain a quantity comparison result, the method further includes:

[0097] If the result of the quantity comparison shows that the number of model combinations in the combined model group set does not meet the preset quantity requirement, then use the target combined model in the combined model group set as the risk control model in the risk control model set, and re-execute the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain the combined model group set until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtain the global combined model set.

[0098] Among them, according to the target combined model, the number of model combinations of each combined model in the target combined model can be determined. If the number of model combinations is greater than or equal to the number of model combinations required in the risk control strategy information, it can be determined that the current target combined model is the global combined model set for screening the final target risk control combined model; if the number of model combinations is less than the number of model combinations required in the risk control strategy information, then it is necessary to perform another permutation and combination on each combined model in the target combined model. The target combined model can be regarded as the initial risk control model set, and each combined model in the target combined model can be regarded as a new model, and perform permutation and combination on each new model until the number of model combinations of the finally obtained combined model is greater than or equal to the number of model combinations required in the risk control strategy information, then the finally obtained combined model set can be determined as the final global combined model set.

[0099] S104. Use the target combined model in the combined model group set as the risk control model in the risk control model set, and re-execute the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain the combined model group set until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtain the global combined model set.

[0100] Among them, when the models required for target business risk control are face recognition model, speech recognition model, and semantic understanding model, after model permutation and combination and comparison with the requirements of risk control strategies, the first target model combination, the second target model combination, and the third target model combination can be obtained. At this time, the number of models in each target model combination is two. If the quantity requirement of the risk control strategy for the target business is three, then the first target model combination, the second target model combination, and the third target model combination need to be regarded as each risk control model in the initially determined risk control model set, and pairwise series permutation and combination are performed on the above first target model combination, second target model combination, and third target model combination, so as to obtain the first updated target model combination between the first target model combination and the second target model combination, the second updated target model combination between the second target model combination and the third target model combination, and the third updated target model combination between the first target model combination and the third target model combination. At this time, the number of model combinations in the first updated target model combination, the second updated target model combination, and the third updated target model combination meets the quantity requirement in the above risk control strategy. Therefore, the first updated target model combination, the second updated target model combination, and the third updated target model combination constitute the global combined model set.

[0101] S105. Determine the target risk control combined model according to the global combined model set and the risk control strategy information.

[0102] Among them, the global combined model set refers to the set of combined models that meet the quantity requirement of the risk control strategy.

[0103] Among them, when the models required for target business risk control are face recognition model, speech recognition model, and semantic understanding model, and the quantity requirement of the risk control strategy is three at the same time, the obtained global combined model set includes the first updated target model set, the second updated target model set, and the third updated target model set. Comparing the above three updated target model sets with the risk control strategy requirements in the risk control strategy information respectively, the first risk control combined model, the second risk control combined model, and the third risk control combined model that meet the risk control strategy requirements can be obtained in the above three updated target model sets. Comparing the above first risk control combined model, the second risk control combined model, and the third risk control combined model with each other, the target risk control combined model with relatively better risk control effect can be determined among the above three risk control combined models.

[0104] Among them, in the embodiment of the present application, determining the target risk control combined model according to the global combined model set and the risk control strategy information includes:

[0105] Determine each combined model in the global combined model set and the risk control effect of each combined model according to the global combined model set;

[0106] Compare the risk control effects of each combined model with the risk control strategy requirements in the risk control strategy information to determine the target risk control combined model that meets the risk control strategy requirements.

[0107] Among them, in the embodiment of the present application, for step S101, the risk control models in the risk control model set are not limited to a single model, and may also include a certain combined model. This combined model can be used as a whole model for subsequent model permutations and combinations, so as to obtain the final target risk control combined model that meets the risk control strategy requirements, which proves that the business risk control system proposed in the embodiment of the present application can not only pre-deploy a single model in advance, but also pre-deploy a combined model, improving the usability of the business risk control system.

[0108] The method for determining the business risk control model provided by the present application can determine the risk control strategy information of the target business and the risk control model set corresponding to the risk control strategy in the risk control strategy information; perform permutation and combination processing on the risk control models in the risk control model set to obtain a combined model group set, the combined model group set includes combined model groups, and the combined model groups include the combined models obtained by performing permutation and combination processing on the risk control models; determine the target combined model in the combined model group set, the target combined model is the combined model that meets the risk control strategy requirements in the combined model group; use the target combined model in the combined model group set as the risk control model in the risk control model set, and re-execute the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain the combined model group set until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtain the global combined model set; determine the target risk control combined model according to the global combined model set and the risk control strategy information, thereby solving the problem that the risk control strategy of the existing risk control products is relatively fixed and cannot solve the business risk control with complex logic.

[0109] Figure 2 It is a schematic flowchart of another method for determining the business risk control model provided by the embodiment of the present application. As Figure 2 shown, the method includes:

[0110] S201. Determine the business requirements of the target business and the target models that meet the business requirements;

[0111] S202. Unifiedly manage the parameter data and training data of the target models.

[0112] Among them, according to the business requirements of the target business, determine the full set of target models that can meet the business requirements and can complete business risk control, and manage the parameter data, business prediction threshold data, training data, verification data, etc. of the full set of target models.

[0113] S203. Based on the AI intelligent platform, model training, parameter tuning, result verification, and model finalization are achieved.

[0114] Among them, the AI intelligent platform has finalized the main models in natural language processing such as text analysis, semantic parsing, sentiment analysis, machine translation, and information extraction; in computer vision, it has finalized models such as image recognition, object detection, face recognition, image generation, account opening business, and change business; in time series prediction, it has finalized models such as threshold prediction, business volume prediction, and sales prediction.

[0115] S204. According to business requirements, use the process concatenation engine to perform matrix arrangement on the standard models that have completed model training to obtain the target combined model.

[0116] Among them, for the standard models that have completed model training and passed verification, according to business requirements, use the process concatenation engine to form specific risk control prediction capabilities. Use the matrix process concatenation engine to arrange the models in a matrix and hybrid arrangement, and realize the concatenation of the models according to the weight algorithm. Finally, form a relatively good ability abstraction of the model.

[0117] Among them, the matrix process concatenation engine abstracts the concatenation of models into a pairwise concatenation method, and performs real-time effect analysis on the concatenation effect. Select the relatively good combination and record the combination result. Finally, recursively concatenate the combination results of pairwise concatenation until the relatively good full-scale concatenation result is selected. Among them, the relatively good combination judgment in the effect analysis can be based on weights or on the way of accuracy verification. Among them, the judgment indicators can be customized and existing standards can be used.

[0118] S205. Display the risk control results and model data in real time.

[0119] Among them, to make the capabilities intuitively visible, establish a large-screen standard graphical display screen for real-time large-screen display of risk control results, construct a medium-screen automated multi-dimensional display report, and realize the real-time warning ability of the small screen, which can ensure the timely discovery of risk control problems. At the same time, the display view also covers the real-time display of internal data of the platform, including data, process data, basic model training data, etc., and provides upload and export capabilities.

[0120] S206. Perform authentication management on registered users, including functions such as ability authentication, transmission encryption, identity authentication, permission management, permission review, role assignment, and data control.

[0121] Among them, to ensure data security, all systems or users that access the risk control platform for ability abstraction or reuse of existing capabilities need to be uniformly and soundly managed on the ability outsourcing platform, and can use the capabilities only after registration is completed.

[0122] Among them, when the business risk control platform is applied to the operator user account opening, the overall business risk control process of the platform is as follows: When handling the account opening business at the front end, the business hall staff requires the applicant to handle it in person and conducts face recognition and live body verification operations. However, it does not rule out the situation where relevant verifications are bypassed due to system problems or other human factors, or the situation where others handle it on behalf of the applicant; there may be unreasonable situations in the application of the tariff packages for business handling during the account opening process; whether the business hall staff clearly informs and whether the customer clearly knows in the notice of business handling instructions are all key points of risk control. The intervention method of the risk control platform is non-intrusive cross-sectional data acquisition. The risk control platform obtains business handling data in real time, including business data, user information-related data, face comparison data, and business hall voice conversation data. After preprocessing such as merging and splitting the data, it flows into the account opening risk control capabilities abstracted by the concatenation of model capabilities. After the data passes through the account opening risk control capabilities, it will predict the legality of the business order data, predict the clarity, identity, and document matching of the face data, and check the voice data between the salesperson and the customer to parse out the informed information, the customer's reply information, and the customer's emotional information to predict the customer's awareness situation, forming the final risk control result, and giving real-time warnings on the small screen for the business handling orders with problems.

[0123] The method for determining the business risk control model provided by the embodiments of the present application establishes a set of funds and business risk control platforms based on the operator system, provides real-time and batch risk control capabilities, is used to prevent the occurrence of risk problems caused by handling high-risk businesses. At the same time, a matrix process concatenation engine is used to realize the arrangement and screening of the models, and all the data of the models are uniformly displayed, ensuring the security of the model data, realizing effective risk control in the business field with complex logic, and promoting the safe and rapid development of the business.

[0124] Figure 3 It is a schematic structural diagram of the device for determining the business risk control model provided by the embodiments of the present application. As Figure 3 shown, the device 30 for determining the business risk control model includes: a policy information determination module 301, a model set obtaining module 302, a combined model determination module 303, a combined model set obtaining module 304, and a target model determination module 305. Among them:

[0125] The policy information determination module 301 is used to determine the risk control policy information of the target business and the risk control model set corresponding to the risk control policy in the risk control policy information;

[0126] The model set obtaining module 302 is used to perform permutation and combination processing on the risk control models in the risk control model set to obtain a combined model set. The combined model set includes combined model groups, and the combined model groups include combined models obtained by performing permutation and combination processing on the risk control models;

[0127] The combined model determination module 303 is configured to determine a target combined model in the combined model group set, where the target combined model is a combined model in the combined model group that meets the requirements of the risk control strategy;

[0128] The combined model set obtaining module 304 is configured to use the target combined model in the combined model group set as the risk control model in the risk control model set, and re - execute the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain the combined model group set, until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtain the global combined model set;

[0129] The target model determination module 305 is configured to determine a target risk control combined model according to the global combined model set and the risk control strategy information.

[0130] In the embodiment of the present application, the policy information determination module 301 may further be specifically configured to:

[0131] Determine the target business and the business type of the target business;

[0132] According to the business type of the target business, determine the risk control strategy information of the target business;

[0133] According to the risk control strategy information, determine the risk control model set corresponding to the risk control strategy in the risk control strategy information.

[0134] In the embodiment of the present application, the model group set obtaining module 302 may further be specifically configured to:

[0135] According to the risk control strategy information of the target business, determine the preset number of models for permutation and combination;

[0136] According to the preset number of models, perform permutation and combination on the risk control models in the risk control model set to obtain each combined model group;

[0137] According to each combined model group, obtain the combined model group set.

[0138] In the embodiment of the present application, the combined model determination module 303 may further be specifically configured to:

[0139] In the embodiment of the present application, determining the target combined model in the combined model group set includes:

[0140] According to the combined model group set, determine the risk control effects of each combined model group in the combined model group set;

[0141] According to the risk control strategy information of the target business, determine the risk control strategy requirements in the risk control strategy information;

[0142] Based on the risk control effects of each combined model group and the requirements of the risk control strategy, the target combined model that meets the requirements of the risk control strategy is obtained from the combined model group set.

[0143] In the embodiment of the present application, the combined model determination module 303 may also be specifically configured to:

[0144] Based on the risk control effects of each combined model group, determine the risk control effects of each combined model in each combined model group;

[0145] Compare the risk control effects of each combined model with each other to obtain the initial target combined models in each combined model group;

[0146] Based on the initial target combined models in each combined model group, determine the target combined model.

[0147] In the embodiment of the present application, the combined model set obtaining module 304 may also be specifically configured to:

[0148] Based on the target combined model, determine the number of model combinations in the target combined model;

[0149] Based on the number of risk control strategies in the risk control strategy information, determine the preset number requirement for the model combination;

[0150] Compare the number of model combinations with the preset number requirement to obtain a number comparison result;

[0151] If the number comparison result is that the number of model combinations meets the preset number requirement, then obtain the global combined model set.

[0152] In the embodiment of the present application, the combined model set obtaining module 304 may also be specifically configured to:

[0153] If the number comparison result is that the number of model combinations does not meet the preset number requirement, then use the target combined model in the combined model group set as the risk control model in the risk control model set, and re - execute the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain the combined model group set until the number of model combinations of the combined models in the obtained combined model group set meets the preset number requirement, and then obtain the global combined model set.

[0154] In the embodiment of the present application, the target model determination module 305 may also be specifically configured to:

[0155] Based on the global combined model set, determine each combined model in the global combined model set and the risk control effect of each combined model;

[0156] Compare the risk control effects of each combined model with the requirements of the risk control strategies in the risk control strategy information to determine the target risk control combined model that meets the requirements of the risk control strategy.

[0157] As can be seen from the above, the determination device 30 of the service risk control model in the embodiment of the present application includes a policy information determination module 301, which is used to determine the risk control policy information of the target service and the risk control model set corresponding to the risk control policy in the risk control policy information; a model group set obtaining module 302, which is used to perform permutation and combination processing on the risk control models in the risk control model set to obtain a combined model group set, the combined model group set includes combined model groups, and the combined model groups include combined models obtained by performing permutation and combination processing on the risk control models; a combined model determination module 303, which is used to determine the target combined model in the combined model group set, and the target combined model is a combined model in the combined model group that meets the risk control policy requirements; a combined model set obtaining module 304, which is used to use the target combined model in the combined model group set as the risk control model in the risk control model set, and re-perform the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain a combined model group set until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtain a global combined model set; a target model determination module 305, which is used to determine a target risk control combined model according to the global combined model set and the risk control policy information. Thus, by mixing and arranging the models and according to the strategy situation of the model combination, a target risk control combined model that can complete the risk control of business development is finally determined.

[0158] Figure 4 It is a schematic structural diagram of an electronic device provided in an embodiment of the present application. As Figure 4 shown, the electronic device 40 includes:

[0159] The electronic device 40 may include a processor 401 with one or more processing cores, a memory 402 with one or more computer-readable storage media, a communication component 403 and other components. Among them, the processor 401, the memory 402 and the communication component 403 are connected through a bus 404.

[0160] In a specific implementation process, at least one processor 401 executes the computer execution instructions stored in the memory 402, so that at least one processor 401 executes the above method for determining the service risk control model.

[0161] For the specific implementation process of the processor 401, reference may be made to the above method embodiment, and its implementation principle and technical effects are similar, which will not be elaborated here in this embodiment.

[0162] In the above Figure 4In the illustrated embodiments, it should be understood that the processor may be a central processing unit (CPU), or it may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in connection with the invention may be directly implemented by a hardware processor or may be implemented by a combination of hardware and software modules in the processor.

[0163] The memory may include a random access memory (RAM), and may also include non-volatile memory (NVM), such as at least one disk memory.

[0164] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, the buses in the drawings of the present application are not limited to only one bus or one type of bus.

[0165] In some embodiments, a computer program product is also proposed, including a computer program or instruction, which when executed by a processor implements the steps in any of the above-mentioned methods for determining a business risk control model.

[0166] For the specific implementation of each of the above operations, reference may be made to the previous embodiments and will not be elaborated herein.

[0167] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions or by controlling relevant hardware through instructions. The instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0168] Therefore, an embodiment of the present application provides a computer-readable storage medium, which stores multiple instructions that can be loaded by a processor to execute the steps in any of the methods for determining a business risk control model provided by the embodiments of the present application.

[0169] Among them, the storage medium may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk, etc.

[0170] According to one aspect of the present application, there is provided a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium.

[0171] Since the instructions stored in the storage medium can execute the steps in any of the methods for determining a service risk control model provided by the embodiments of the present application, the beneficial effects achievable by any of the methods for determining a service risk control model provided by the embodiments of the present application can be realized. For details, see the previous embodiments and will not be repeated here.

[0172] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any variations, uses, or adaptations of the present application, which follow the general principles of the present application and include known common knowledge or conventional technical means in the technical field not disclosed in the present application. The specification and embodiments are only regarded as exemplary, and the true scope and spirit of the present application are pointed out by the following claims.

[0173] It should be understood that the present application is not limited to the exact structures already described and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. A method for determining a business risk control model, characterized in that, the method includes: determining risk control strategy information of a target business and a set of risk control models corresponding to the risk control strategies in the risk control strategy information; performing permutation and combination processing on the risk control models in the set of risk control models to obtain a set of combined model groups, the set of combined model groups includes combined model groups, and the combined model groups include combined models obtained by performing permutation and combination processing on the risk control models; determining a target combined model in the set of combined model groups, the target combined model being a combined model in the combined model group that meets the requirements of the risk control strategy; using the target combined model in the set of combined model groups as the risk control model in the set of risk control models, and re-performing the step of performing permutation and combination processing on the risk control models in the set of risk control models to obtain a set of combined model groups until the number of model combinations of the combined models in the obtained set of combined model groups meets the preset quantity requirement, and then obtaining a global set of combined models; determining a target risk control combined model according to the global set of combined models and the risk control strategy information.

2. The method according to claim 1, characterized in that, the determining of the risk control strategy information of the target business and the set of risk control models corresponding to the risk control strategies in the risk control strategy information includes: determining a target business and the business type of the target business; determining the risk control strategy information of the target business according to the business type of the target business; determining a set of risk control models corresponding to the risk control strategies in the risk control strategy information according to the risk control strategy information.

3. The method according to claim 1, characterized in that, the performing of permutation and combination processing on the risk control models in the set of risk control models to obtain a set of combined model groups includes: determining a preset number of models for permutation and combination according to the risk control strategy information of the target business; performing permutation and combination on the risk control models in the set of risk control models according to the preset number of models to obtain each combined model group; obtaining a set of combined model groups according to each of the combined model groups.

4. The method according to claim 1, characterized in that, the determining of the target combined model in the set of combined model groups includes: determining the risk control effects of each combined model group in the set of combined model groups according to the set of combined model groups; determining the requirements of the risk control strategies in the risk control strategy information according to the risk control strategy information of the target business; obtaining the target combined model in the set of combined model groups that meets the requirements of the risk control strategy according to the risk control effects of each combined model group and the requirements of the risk control strategy.

5. The method according to claim 4, characterized in that, after the determining of the risk control effects of each combined model group in the set of combined model groups according to the set of combined model groups, the method further includes: determining the risk control effects of each combined model in each combined model group according to the risk control effects of each combined model group; comparing the risk control effects of each combined model with each other to obtain the initial target combined model in each combined model group; Determine the target combined model according to the initial target combined models in each of the described combined model groups.

6. The method according to claim 1, wherein, after determining the target combined model in the combined model group set and before using the target combined model in the combined model group set as the risk control model in the risk control model set, the method further includes: Determine the number of model combinations in the target combined model according to the target combined model; Determine the preset quantity requirement of the model combination according to the number of risk control strategies in the risk control strategy information; Compare the number of model combinations with the preset quantity requirement to obtain a quantity comparison result; If the quantity comparison result is that the number of model combinations meets the preset quantity requirement, then obtain the global combined model set.

7. The method according to claim 6, wherein, after comparing the number of model combinations with the preset quantity requirement to obtain a quantity comparison result, the method further includes: If the quantity comparison result is that the number of model combinations does not meet the preset quantity requirement, then use the target combined model in the combined model group set as the risk control model in the risk control model set, and re - execute the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain the combined model group set, until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtain the global combined model set.

8. The method according to claim 1, wherein, the determining the target risk control combined model according to the global combined model set and the risk control strategy information includes: Determine each combined model in the global combined model set and the risk control effect of each combined model according to the global combined model set; Compare the risk control effects of each combined model with the risk control strategy requirements in the risk control strategy information to determine the target risk control combined model that meets the risk control strategy requirements.

9. An apparatus for determining a business risk control model, wherein, comprises: A strategy information determination module, configured to determine the risk control strategy information of the target business and the risk control model set corresponding to the risk control strategies in the risk control strategy information; A model group set obtaining module, configured to perform permutation and combination processing on the risk control models in the risk control model set to obtain a combined model group set, the combined model group set includes combined model groups, and the combined model groups include combined models obtained by performing permutation and combination processing on the risk control models; A combined model determination module, configured to determine the target combined model in the combined model group set, and the target combined model is the combined model in the combined model group that meets the risk control strategy requirements; A combined model set obtaining module, configured to use the target combined model in the combined model set as the risk control model in the risk control model set, and re - execute the step of performing permutation and combination processing on the risk control models in the risk control model set to obtain the combined model group set, until the number of model combinations of the combined models in the obtained combined model group set meets the preset quantity requirement, and then obtain the global combined model set; A target model determination module, configured to determine a target risk control combination model according to the global combination model set and the risk control policy information.

10. An electronic device, characterized in that it includes: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the method according to any one of claims 1 to 8.

Images