Risk assessment method and device, equipment and storage medium

Abstract

The invention discloses a risk assessment method, device and equipment and a storage medium, which are applied to financial science and technology scenes, and the method comprises the steps: carrying out the gray wolf optimization algorithm initialization of the features of a loan user through employing a chaotic mapping and opposite learning strategy, so as to obtain a gray wolf group; an optimal solution, a second optimal solution and a third optimal solution are screened out from the grey wolf group, and when it is determined that the optimal solution meets a variation condition, the optimal solution is varied; obtaining an updated residual solution; obtaining an updated grey wolf group according to the varied optimal solution, the varied second optimal solution, the varied third optimal solution and the updated residual solution, and determining a target solution from the updated grey wolf group when an iteration termination condition is met; and performing risk assessment on the loan user according to the target feature subset corresponding to the target solution. The population diversity is improved through opposite learning, the optimal solution is mutated, and the remaining solution is updated and searched according to the mutated optimal solution, so that the algorithm is prevented from being caught in local optimum too early, and accurate risk assessment of the loan user is realized.

Description

Technical Field

[0001] The present invention relates to the field of financial technology, and in particular to a risk assessment method, device, equipment and storage medium. Background Art

[0002] Financial institutions often need to handle a large number of loan businesses, and they usually need to review the user's loan qualifications to assess the user's loan risk, and take corresponding loan restriction measures when the risk is relatively high to protect the financial institution's funds. When conducting risk assessment, feature extraction algorithms are usually used to screen features from the user's numerous features, and perform risk assessment based on the screened features.

[0003] Currently, the most commonly used method is to use the multi-objective grey wolf optimization algorithm for feature extraction. However, when using the traditional grey wolf optimization algorithm for feature extraction, due to the limitations of the algorithm itself, it is difficult to maintain population diversity and it is easy to fall into local optimality too early, so that the obtained feature solution may not be the optimal solution. Therefore, when conducting risk assessment of loan users based on the feature solution obtained by the above algorithm, the assessment results will be inaccurate. Summary of the invention

[0004] The present invention provides a risk assessment method to achieve accurate risk assessment of loan users.

[0005] According to a first aspect of the present invention, there is provided a risk assessment method, comprising: using a chaotic mapping and adversarial learning strategy to initialize a gray wolf optimization algorithm on loan user characteristics to obtain a gray wolf group, wherein the gray wolf group includes solutions of multiple feature subsets;

[0006] Selecting an optimal solution, a suboptimal solution, and a third optimal solution from the gray wolf group, and mutating the optimal solution when it is determined that the optimal solution meets a mutation condition;

[0007] Performing local search and updating on the remaining solutions after screening according to the optimal solution after mutation to obtain updated remaining solutions;

[0008] Obtaining an updated gray wolf population according to the mutated optimal solution, the suboptimal solution, the third optimal solution, and the updated remaining solutions, and determining a target solution from the updated gray wolf population when an iteration termination condition is met;

[0009] The risk of the loan user is assessed according to the target feature subset corresponding to the target solution.

[0010] According to another aspect of the present invention, there is provided a risk assessment device, comprising: an initialization module, for initializing the characteristics of loan users using a gray wolf optimization algorithm using a chaotic mapping and adversarial learning strategy to obtain a gray wolf group, wherein the gray wolf group includes solutions for multiple characteristic subsets;

[0011] A mutation module, used for selecting an optimal solution, a suboptimal solution and a third optimal solution from the gray wolf group, and mutating the optimal solution when it is determined that the optimal solution meets a mutation condition;

[0012] A local update module is used to perform local search and update on the remaining solutions after screening according to the optimal solution after mutation to obtain updated remaining solutions;

[0013] A target solution determination module is used to obtain an updated gray wolf population according to the mutated optimal solution, the suboptimal solution, the third optimal solution and the updated remaining solutions, and determine a target solution from the updated gray wolf population when an iteration termination condition is met;

[0014] The risk assessment module is used to perform risk assessment on the loan user according to the target feature subset corresponding to the target solution.

[0015] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0016] at least one processor; and

[0017] a memory communicatively connected to the at least one processor; wherein,

[0018] The memory stores a computer program that can be executed by the at least one processor. The computer program is executed by the at least one processor so that the at least one processor can perform the method described in any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method described in any embodiment of the present invention when executed.

[0020] According to another aspect of the present invention, a computer program product is provided, comprising a computer program, and when the computer program is executed by a processor, the method described in any embodiment of the present invention is implemented.

[0021] The beneficial technical effect of the present invention is that the initial solution is determined by combining chaotic mapping with adversarial learning, so that the distribution of the initial solution is more uniform and the population diversity is improved. When it is determined that the mutation conditions are met, the solution is mutated and the remaining solutions are updated and searched with reference to the optimal solution after the mutation, so as to avoid the algorithm from falling into the local optimum too early, so that the target solution finally obtained is more accurate, thereby realizing accurate risk assessment of loan users based on the target solution.

[0022] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 is a flow chart of a risk assessment method provided according to Embodiment 1 of the present invention;

[0025] Figure 2 is a schematic diagram of a data structure of a solution provided according to Embodiment 1 of the present invention;

[0026] Figure 3 is a flow chart of a risk assessment method provided according to Embodiment 2 of the present invention;

[0027] Figure 4 is a schematic diagram of the structure of a risk assessment device provided according to Embodiment 3 of the present invention;

[0028] Figure 5 It is a schematic diagram of the structure of an electronic device provided by Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 should fall within the scope of protection of the present invention.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] Embodiment 1

[0032] Figure 1 A flow chart of a risk assessment method is provided for the first embodiment of the present invention. This embodiment is applicable to the case of risk assessment of loan users. The method can be executed by a risk assessment device, which can be implemented in the form of hardware and / or software. Figure 1 As shown, the method includes:

[0033] Step S101, using the strategy of chaotic mapping and adversarial learning to initialize the gray wolf optimization algorithm on the characteristics of the loan users to obtain a gray wolf group.

[0034] Optionally, a chaotic mapping and adversarial learning strategy is used to initialize the gray wolf optimization algorithm on the characteristics of loan users to obtain a gray wolf group, including: using a chaotic mapping strategy to initialize the gray wolf optimization algorithm on the characteristics of loan users to obtain an original solution; using an adversarial learning strategy to calculate the original solution to obtain an adversarial solution corresponding to the original solution; combining the original solution and the adversarial solution to construct a gray wolf group.

[0035] Among them, in this implementation, the characteristics of the loan user are obtained according to the registration information filled in by the loan user, wherein the characteristics include age, gender, loan term, etc. Of course, this implementation is only an example, and does not limit the specific type of characteristics. As long as it is related to risk assessment, it is within the scope of protection of this application. In addition, the information collected in this implementation is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data comply with the relevant laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0036] Specifically, since there are many features related to loan users, it is necessary to select a subset that is predictive or best represents the data features for risk assessment. In this implementation, an improved multi-objective gray wolf optimization algorithm is used to select features. When the improved multi-objective gray wolf optimization algorithm is used for population initialization, an Opposition-Based Learning (OBL) strategy is introduced. The OBL strategy is to find the optimal solution from two directions. It can be combined with other strategies, such as chaotic mapping. First, the original solution is generated by the selected other strategies, and then the OBL strategy is used to generate the original solution's opposite solution, so that the distribution of the initial solution composed of the original solution and the opposite solution is more uniform, thereby improving the diversity of the population.

[0037] Among them, the initial solution composed of the original solution and the opposite solution is the gray wolf group, and the gray wolf group includes solutions of multiple feature subsets, such as Figure 2 As shown in the figure, it is a schematic diagram of the data structure of the solution. The solution to the feature selection problem of the multi-objective gray wolf optimization algorithm is divided into two parts. The first part is the feature vector part Ndim, and the second part is the classifier parameter vector part Nparm. For the Ndim part, the value of each feature bit is 1 or 0, 1 represents the selection of the feature, and 0 represents the non-selection of the feature. Therefore, the solution space of the first part can be regarded as a binary discrete solution space. For the Nparm part, there are three digits in total. The first digit is kernel, which represents the selection of the kernel function. 1 represents selection and 0 represents non-selection. The second digit is parameter c, which represents the penalty coefficient, which is in the range of [2 -1 , 2 5 ]; The third is the parameter gamma. From this value, we can see the law of data after mapping the data to a higher-dimensional feature space. The range of gamma is [2 -4 , 2 5 ]. Therefore, a solution to the feature selection problem can be obtained by the one-dimensional array Of course, this embodiment is only an example and does not limit the data structure of the solution.

[0038] In a specific implementation, when the gray wolf optimization algorithm is initialized for the loan user's features through chaotic mapping to obtain 15 original solutions, and the original solutions are calculated using the strategy of adversarial learning, 15 adversarial solutions corresponding to the original solutions are obtained. For example, when one of the original solutions is [0 1 0 1 0 0 1 0 12 10], the Ndim part is completely negated, and an adversarial solution obtained is [1 0 1 0 1 1 0 0 12 10]. Of course, this embodiment is only an example, and the method for obtaining other adversarial solutions is roughly the same, and will not be repeated in this embodiment. Therefore, after the above-mentioned chaotic mapping and adversarial learning, there are a total of 30 solutions in the gray wolf group initially obtained. Of course, this embodiment is only an example, and the number of solutions contained in the gray wolf group is not limited.

[0039] It should be noted that in this embodiment, after the gray wolf group is obtained through the above-mentioned initialization strategy, the specified parameters of the gray wolf optimization algorithm itself will also be initialized. This embodiment does not limit the specific types of parameters involved in the algorithm.

[0040] Step S102, selecting the optimal solution, the suboptimal solution and the third optimal solution from the gray wolf group, and mutating the optimal solution when it is determined that the optimal solution meets the mutation condition.

[0041] Optionally, the optimal solution, the suboptimal solution and the third optimal solution are screened out from the gray wolf group, including: calculating the objective function value corresponding to each solution in the gray wolf group; screening out non-dominated solutions that meet specified conditions from the gray wolf group according to the objective function value; when it is determined that the current number of iterations is less than a preset total number of iterations, screening out the optimal solution, the suboptimal solution and the third optimal solution from the non-dominated solutions.

[0042] Optionally, when it is determined that the optimal solution meets the mutation conditions, the optimal solution is mutated, including: calculating a dynamic threshold corresponding to the current number of iterations based on the current number of iterations and the total number of iterations; obtaining a random number generated at the current number of iterations, and the number of features included in the optimal solution; when the random number is less than the dynamic threshold and the number of features is less than the specified number of features, the feature selection in the optimal solution is inverted to obtain the mutated optimal solution.

[0043] Specifically, in this embodiment, after the gray wolf group is obtained by initialization, each solution will be substituted into the preset multi-objective function, and the objective function value corresponding to each solution will be obtained, so as to screen out non-dominated solutions that meet the specified conditions from the gray wolf group based on the objective function value, wherein the specified conditions may be that the number of features is as small as possible and the classification accuracy is as high as possible. For example, 6 solutions that meet the requirements of the number of features and classification accuracy are obtained from the above 30 solutions through screening, and the above 6 solutions are taken as non-dominated solutions. Of course, this embodiment is only an example, and the specific number of non-dominated solutions screened is not limited. And when it is determined that the current number of iterations t is less than the preset total number of iterations maxiter, the optimal solution α, the suboptimal solution β and the third optimal solution ω will be selected from the non-dominated solutions according to the crowding distance of the wolf group.

[0044] It should be noted that, in this implementation, when the optimal solution α is far away from the actual optimal solution or falls into a local optimum, the performance of the algorithm may be greatly reduced. Therefore, in this implementation, the following formula (1) is used to set a dynamic threshold φ:

[0045] φ=0.5*t / maxiter (1)

[0046] Among them, φ represents the dynamic threshold, t represents the current number of iterations, and maxiter represents the total number of iterations. Since the value of t changes for each iteration, φ also changes dynamically with t.

[0047] Since a random number Nrand is generated for each iteration process, the following formula (2) is used to perform mutation operation on the optimal solution α:

[0048]

[0049] in, Represents the feature of the j-th feature bit in the optimal solution α vector. When it is determined that the random number Nrand is less than the dynamic threshold φ, and the number of feature bits contained in the feature vector part is less than 60, it is necessary to invert the feature on the feature bit, otherwise the feature on the feature bit remains unchanged. Of course, this implementation is only an example and does not limit the specific variation process of the optimal solution.

[0050] Step S103, performing local search and update on the remaining solutions after screening according to the mutated optimal solution to obtain updated remaining solutions.

[0051] Optionally, the remaining solutions after screening are locally searched and updated according to the optimal solution after mutation to obtain updated remaining solutions, including: determining the target direction according to the optimal solution after mutation, searching the remaining solutions according to the target direction to update the feature vector part of the remaining solutions; screening out half of the high-quality solutions from the remaining solutions according to the specified evaluation criteria, and updating the classifier parameter vector part of the high-quality solutions.

[0052] Specifically, in this embodiment, when the optimal solution α mutates, the position of the mutated optimal solution α will change accordingly. Since each solution in the Grey Wolf Optimization Algorithm changes following the optimal solution α, the target direction will be re-determined according to the mutated optimal solution, so that the remaining solutions are re-searched according to the newly determined target direction, thereby increasing the probability of finding a new optimal solution, and in the later stage of the iteration, the feature vector part of the remaining solutions will be updated accordingly after the re-search.

[0053] It should be noted that in order not to increase the time complexity of the calculation, in this implementation, half of the better solutions will be screened out from the remaining solutions, and the parameters c and gamma in the classifier parameter vector part of the better solutions will be updated. Of course, this implementation is only an example and does not limit the specific update process of the remaining solutions.

[0054] It is worth mentioning that in this embodiment, when the solutions in the initially obtained gray wolf group change accordingly, the specified parameters of the gray wolf optimization algorithm itself will be updated again, so that the parameters of the algorithm itself and the changes in the solutions occur in correspondence. Of course, this embodiment does not limit the specific update results of the specified parameters.

[0055] Step S104, obtaining an updated gray wolf population according to the mutated optimal solution, the suboptimal solution, the third optimal solution and the updated remaining solutions, and determining a target solution from the updated gray wolf population when an iteration termination condition is met.

[0056] Optionally, when the iteration termination condition is met, a target solution is determined from the updated gray wolf group, including: calculating the objective function value corresponding to each solution in the updated gray wolf group; screening out updated non-dominated solutions that meet specified conditions from the updated gray wolf group according to the objective function value; when it is determined that the number of updated non-dominated solutions is greater than the required number, screening out the target solution from the updated non-dominated solutions, wherein each target solution corresponds to a different optimization objective.

[0057] Specifically, in this implementation, the mutated optimal solution, suboptimal solution, third optimal solution and the updated remaining solutions will be re-acquired to update the gray wolf group, and the updated non-dominated solution will be screened out from the updated gray wolf group again according to the target. The screening method for updating the non-dominated solution is roughly the same as the non-dominated solution mentioned above, so it will not be repeated in this implementation.

[0058] It should be noted that the required number of non-dominated solutions is pre-set in this embodiment, for example, 6. When it is determined that the number of updated non-dominated solutions obtained is less than the required number, the iteration continues. When the number of updated non-dominated solutions is not less than the required number, the iteration is terminated and the target solution is screened out from the updated non-dominated solutions. For example, when the number of updated non-dominated solutions is exactly 6, all updated non-dominated solutions are used as target solutions; when the number of updated non-dominated solutions is 8, 6 better updated non-dominated solutions are screened out as target solutions. The specific principle of screening can be set according to actual needs, and the specific screening method of the target solution is not limited in this embodiment.

[0059] Step S105, performing risk assessment on the loan user according to the target feature subset corresponding to the target solution.

[0060] Optionally, risk assessment of loan users is performed based on a target feature subset corresponding to the target solution, including: receiving a risk assessment instruction, wherein the risk assessment instruction includes a selected optimization target; determining a matching target solution corresponding to the selected optimization target, and using the matching target solution to perform risk assessment on loan users.

[0061] Specifically, in this embodiment, since there are multiple target solutions obtained, and each target solution corresponds to a different optimization goal, for example, the optimization goal corresponding to target solution 1 is fewer features, and the optimization goal corresponding to target solution 2 is high classification accuracy. Of course, this embodiment is only an example, and does not limit the specific type of optimization goal, which can be set according to the specific needs of risk assessment. And bank staff can select a group of target solutions for risk assessment based on risk assessment needs.

[0062] For example, when the optimization goal included in the received risk assessment instruction is high classification accuracy, target solution 2 is screened out from all the target solutions obtained, and target solution 2 includes four features of age, gender, loan amount and years of work experience, then the terminal device will perform risk assessment on the loan user based on the above four features in target solution 2. In particular, when performing risk assessment, different features can be normalized to obtain a normalized value, and the weight corresponding to each feature can be obtained, so as to calculate the product sum of the normalized value of each feature and the corresponding weight. When the obtained product sum meets the pre-set risk threshold, the loan user is determined to be a safe user, otherwise, the loan user is determined to be a risky user, and in this embodiment, the risk level of the risky user can also be determined according to the specific value of the product sum. Of course, this embodiment is only an example, and does not limit the specific method of risk assessment.

[0063] In the implementation manner of the present application, the initial solution is determined by combining chaotic mapping with adversarial learning, so that the distribution of the initial solution is more uniform and the population diversity is improved. When it is determined that the mutation conditions are met, the solution is mutated and the remaining solutions are updated and searched with reference to the optimal solution after the mutation, so as to avoid the algorithm from falling into the local optimum too early, and the target solution finally obtained is more accurate, thereby achieving accurate risk assessment of loan users based on the target solution.

[0064] Embodiment 2

[0065] Figure 3 This is a flow chart of a risk assessment method provided by Embodiment 2 of the present invention. This embodiment is based on the above embodiment and, after performing risk assessment on loan users according to the target feature subset corresponding to the target solution, further includes: verifying the risk assessment results of loan users. Figure 3 As shown, the method includes:

[0066] Step S201, using the strategy of chaotic mapping and adversarial learning to initialize the gray wolf optimization algorithm on the characteristics of the loan users to obtain a gray wolf group.

[0067] Optionally, a chaotic mapping and adversarial learning strategy is used to initialize the gray wolf optimization algorithm on the characteristics of loan users to obtain a gray wolf group, including: using a chaotic mapping strategy to initialize the gray wolf optimization algorithm on the characteristics of loan users to obtain an original solution; using an adversarial learning strategy to calculate the original solution to obtain an adversarial solution corresponding to the original solution; combining the original solution and the adversarial solution to construct a gray wolf group.

[0068] Step S202, selecting the optimal solution, the suboptimal solution and the third optimal solution from the wolf group, and mutating the optimal solution when it is determined that the optimal solution meets the mutation condition.

[0069] Optionally, the optimal solution, the suboptimal solution and the third optimal solution are screened out from the gray wolf group, including: calculating the objective function value corresponding to each solution in the gray wolf group; screening out non-dominated solutions that meet specified conditions from the gray wolf group according to the objective function value; when it is determined that the current number of iterations is less than a preset total number of iterations, screening out the optimal solution, the suboptimal solution and the third optimal solution from the non-dominated solutions.

[0070] Step S203, locally search and update the remaining solutions after screening according to the mutated optimal solution to obtain updated remaining solutions.

[0071] Optionally, the remaining solutions after screening are locally searched and updated according to the optimal solution after mutation to obtain updated remaining solutions, including: determining the target direction according to the optimal solution after mutation, searching the remaining solutions according to the target direction to update the feature vector part of the remaining solutions; screening out half of the high-quality solutions from the remaining solutions according to the specified evaluation criteria, and updating the classifier parameter vector part of the high-quality solutions.

[0072] Step S204, obtaining an updated gray wolf population according to the mutated optimal solution, the suboptimal solution, the third optimal solution and the updated remaining solutions, and determining a target solution from the updated gray wolf population when an iteration termination condition is met.

[0073] Optionally, when the iteration termination condition is met, a target solution is determined from the updated gray wolf group, including: calculating the objective function value corresponding to each solution in the updated gray wolf group; screening out updated non-dominated solutions that meet specified conditions from the updated gray wolf group according to the objective function value; when it is determined that the number of updated non-dominated solutions is greater than the required number, screening out the target solution from the updated non-dominated solutions, wherein each target solution corresponds to a different optimization objective.

[0074] Step S205: Perform risk assessment on the loan user according to the target feature subset corresponding to the target solution.

[0075] Optionally, risk assessment of loan users is performed based on a target feature subset corresponding to the target solution, including: receiving a risk assessment instruction, wherein the risk assessment instruction includes a selected optimization target; determining a matching target solution corresponding to the selected optimization target, and using the matching target solution to perform risk assessment on loan users.

[0076] Step S206: verifying the risk assessment result of the loan user.

[0077] Specifically, in this embodiment, after obtaining the risk assessment results of the loan user, the risk assessment results will also be verified. That is, after obtaining the risk assessment results of the loan user, the actual risk assessment results determined by manual evaluation can be received, and the actual risk assessment results can be compared with the predicted risk assessment results. If the risk assessment result automatically predicted is a safe user, but the risk assessment result obtained manually is a high-risk user, then it is determined that the risk assessment result predicted by the above algorithm is wrong.

[0078] It should be noted that the reason for the huge gap between the above two evaluation results may be that the improved Grey Wolf optimization algorithm has an error, or the hardware device itself has an error. The specific cause of the evaluation error is not limited in this implementation. And when the evaluation result obtained by the algorithm is obviously wrong, an alarm message will be generated to prompt the operation and maintenance personnel to promptly detect and maintain the hardware device or software algorithm, so as to ensure the accuracy of subsequent risk assessments of other loan users.

[0079] It is worth mentioning that in order to save operating resources, in this implementation, verification can be performed in a periodic timing manner, that is, a verification request will be sent to the operation and maintenance personnel every other day. When the operation and maintenance personnel receive the verification request, they can accept or reject it. When it is confirmed that it is accepted, a risk assessment is performed on the current loan user to be assessed, and the risk assessment result is compared with the risk assessment result automatically obtained by the algorithm to obtain the verification result, so that when an assessment error occurs, regular maintenance of the algorithm or hardware equipment can be achieved.

[0080] In the implementation manner of the present application, the initial solution is determined by combining chaotic mapping with adversarial learning, so that the distribution of the initial solution is more uniform and the population diversity is improved. When it is determined that the mutation conditions are met, the solution is mutated and the remaining solutions are updated and searched with reference to the optimal solution after the mutation, so as to avoid the algorithm from falling into the local optimum too early, and the target solution finally obtained is more accurate, thereby achieving accurate risk assessment of loan users based on the target solution.

[0081] Embodiment 3

[0082] Figure 4 This is a schematic diagram of the structure of a risk assessment device provided in Embodiment 3 of the present invention. Figure 4 As shown, the device comprises:

[0083] The initialization module 310 is used to use the strategy of chaotic mapping and adversarial learning to initialize the characteristics of the loan user with a gray wolf optimization algorithm to obtain a gray wolf group, wherein the gray wolf group includes solutions of multiple feature subsets;

[0084] A mutation module 320 is used to select an optimal solution, a suboptimal solution and a third optimal solution from the gray wolf group, and mutate the optimal solution when it is determined that the optimal solution meets the mutation condition;

[0085] A local update module 330 is used to perform local search and update on the remaining solutions after screening according to the optimal solution after mutation to obtain updated remaining solutions;

[0086] The target solution determination module 340 is used to obtain an updated gray wolf population according to the mutated optimal solution, the suboptimal solution, the third optimal solution and the updated remaining solutions, and determine the target solution from the updated gray wolf population when the iteration termination condition is met;

[0087] The risk assessment module 350 is used to perform risk assessment on loan users according to the target feature subset corresponding to the target solution.

[0088] Optionally, an initialization module is used to use a chaotic mapping strategy to initialize the gray wolf optimization algorithm on the characteristics of the loan user to obtain an original solution;

[0089] The original solution is calculated using the adversarial learning strategy to obtain the adversarial solution corresponding to the original solution;

[0090] The original solution and the opposing solution are combined to construct the gray wolf population.

[0091] Optionally, the mutation module includes a solution screening unit for calculating the objective function value corresponding to each solution in the gray wolf population;

[0092] According to the objective function value, non-dominated solutions that meet the specified conditions are selected from the gray wolf population;

[0093] When it is determined that the current number of iterations is less than the preset total number of iterations, the optimal solution, the suboptimal solution and the third optimal solution are screened out from the non-dominated solutions.

[0094] Optionally, the mutation module includes a mutation unit, which is used to calculate a dynamic threshold corresponding to the current number of iterations according to the current number of iterations and the total number of iterations;

[0095] Get the random number generated at the current number of iterations and the number of features contained in the optimal solution;

[0096] When the random number is less than the dynamic threshold and the number of features is less than the specified number of features, the feature selection in the optimal solution is inverted to obtain the optimal solution after mutation.

[0097] Optionally, the data structure of the solution includes a feature vector part and a classifier parameter vector part; a local update module, used to determine the target direction according to the optimal solution after mutation, and search the remaining solutions according to the target direction to update the feature vector part of the remaining solutions;

[0098] Half of the high-quality solutions are selected from the remaining solutions according to the specified evaluation criteria, and the classifier parameter vector part of the high-quality solutions is updated.

[0099] Optionally, a target solution determination module is used to calculate and update the target function value corresponding to each solution in the gray wolf group;

[0100] According to the objective function value, the updated non-dominated solution that meets the specified conditions is selected from the updated gray wolf group;

[0101] When it is determined that the number of updated non-dominated solutions is greater than the required number, target solutions are screened out from the updated non-dominated solutions, wherein each target solution corresponds to a different optimization goal.

[0102] Optionally, a risk assessment module is used to receive a risk assessment instruction, wherein the risk assessment instruction includes a selected optimization target;

[0103] Determine the matching target solution corresponding to the selected optimization goal, and use the matching target solution to conduct risk assessment on loan users.

[0104] The risk assessment device provided in the embodiment of the present invention can execute the risk assessment method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0105] Embodiment 4

[0106] Figure 5 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0107] like Figure 5 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0108] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0109] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the risk assessment method.

[0110] In some embodiments, the risk assessment method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the risk assessment method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the risk assessment method in any other appropriate manner (e.g., by means of firmware).

[0111] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0112] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0113] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0114] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0115] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0116] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0117] Embodiment 5

[0118] An embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the risk assessment method provided in any embodiment of the present application.

[0119] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0120] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0121] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A risk assessment method, characterized in that: include: Using the strategy of chaotic mapping and adversarial learning to initialize the characteristics of loan users with a gray wolf optimization algorithm to obtain a gray wolf group, wherein the gray wolf group includes solutions of multiple feature subsets; Selecting an optimal solution, a suboptimal solution, and a third optimal solution from the gray wolf group, and mutating the optimal solution when it is determined that the optimal solution meets a mutation condition; Performing local search and updating on the remaining solutions after screening according to the optimal solution after mutation to obtain updated remaining solutions; Obtaining an updated gray wolf population according to the mutated optimal solution, the suboptimal solution, the third optimal solution, and the updated remaining solutions, and determining a target solution from the updated gray wolf population when an iteration termination condition is met; The risk of the loan user is assessed according to the target feature subset corresponding to the target solution.

2. The method according to claim 1, characterized in that The strategy of using chaotic mapping and adversarial learning to initialize the characteristics of loan users with a gray wolf optimization algorithm to obtain a gray wolf group includes: Using a chaotic mapping strategy to initialize the characteristics of the loan user using a grey wolf optimization algorithm to obtain an original solution; Calculating the original solution using the adversarial learning strategy to obtain an adversarial solution corresponding to the original solution; The original solution and the opposing solution are combined to construct the gray wolf group.

3. The method according to claim 1, characterized in that The step of selecting the optimal solution, the second-optimal solution and the third-optimal solution from the gray wolf group includes: Calculating the objective function value corresponding to each solution in the gray wolf group; Screening out non-dominated solutions that meet specified conditions from the gray wolf population according to the objective function value; When it is determined that the current number of iterations is less than the preset total number of iterations, the optimal solution, the suboptimal solution and the third optimal solution are screened out from the non-dominated solutions.

4. The method according to claim 3, characterized in that When it is determined that the optimal solution satisfies a mutation condition, mutating the optimal solution comprises: Calculating a dynamic threshold corresponding to the current number of iterations according to the current number of iterations and the total number of iterations; Obtaining the random number generated at the current number of iterations and the number of features included in the optimal solution; When the random number is smaller than the dynamic threshold and the number of features is smaller than the specified number of features, the feature selection in the optimal solution is inverted to obtain the optimal solution after mutation.

5. The method according to claim 1, characterized in that The data structure of each solution includes a feature vector part and a classifier parameter vector part; The method of performing local search and updating on the filtered remaining solutions according to the mutated optimal solution to obtain updated remaining solutions includes: Determining a target direction according to the optimal solution after the variation, searching the remaining solutions according to the target direction to update the feature vector part of the remaining solutions; Half of the high-quality solutions are selected from the remaining solutions according to a specified evaluation standard, and the classifier parameter vector part of the high-quality solutions is updated.

6. The method according to claim 3, characterized in that The step of determining a target solution from the updated gray wolf population when an iteration termination condition is met includes: Calculating the objective function value corresponding to each solution in the updated gray wolf population; According to the objective function value, an updated non-dominated solution satisfying the specified condition is selected from the updated gray wolf population; When it is determined that the number of the updated non-dominated solutions is greater than the required number, the target solutions are screened out from the updated non-dominated solutions, wherein each of the target solutions corresponds to a different optimization goal.

7. The method according to claim 1, characterized in that The risk assessment of the loan user according to the target feature subset corresponding to the target solution includes: Receiving a risk assessment instruction, wherein the risk assessment instruction includes a selected optimization objective; A matching target solution corresponding to the selected optimization target is determined, and the matching target solution is used to perform risk assessment on the loan user.

8. A risk assessment device, characterized in that: The device comprises: An initialization module, used to use a chaotic mapping and adversarial learning strategy to initialize the characteristics of loan users using a gray wolf optimization algorithm to obtain a gray wolf group, wherein the gray wolf group includes solutions for multiple feature subsets; A mutation module, used for selecting an optimal solution, a suboptimal solution and a third optimal solution from the gray wolf group, and mutating the optimal solution when it is determined that the optimal solution meets a mutation condition; A local update module is used to perform local search and update on the remaining solutions after screening according to the optimal solution after mutation to obtain updated remaining solutions; A target solution determination module is used to obtain an updated gray wolf population according to the mutated optimal solution, the suboptimal solution, the third optimal solution and the updated remaining solutions, and determine a target solution from the updated gray wolf population when an iteration termination condition is met; The risk assessment module is used to perform risk assessment on the loan user according to the target feature subset corresponding to the target solution.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method according to any one of claims 1 to 7 when executed.

11. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 7 when being executed by a processor.

Images