Credit risk prediction method and device based on particle swarm optimization, equipment and storage medium
Through the combination of support vector machine and particle swarm optimization algorithm, a more accurate credit risk prediction model was generated, which solved the problem of strong subjectivity of existing credit assessment methods, improved the accuracy and efficiency of credit risk prediction, and reduced the bank's Bad debt risk.
Patent Information
- Application Number
- CN202411023488.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-12-06
AI Technical Summary
The existing credit risk assessment methods are subjective and lack quantitative standards, making it difficult to effectively predict the repayment ability and repayment awareness of loan applicants, resulting in banks facing bad debt risks.
The support vector machine method is used to train the credit data, and the kernel function parameters are optimized using the particle swarm optimization algorithm to generate a more accurate credit risk prediction model.
By optimizing the support vector machine model by particle swarm, the accuracy and efficiency of credit risk prediction are improved, the risk of bank bad debts is reduced, and more scientific credit evaluation methods are provided.
Smart Images

Figure CN119090619A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent algorithms, in particular to support vector machine methods and particle swarm optimization algorithms, and more particularly to a credit risk prediction method, device and storage medium based on particle swarm optimization. Background Art
[0002] At present, for commercial banks, developing credit business is an important way to increase bank profit growth points and improve the quality of bank customer service. However, on the other hand, the risks brought by credit business are also important issues that commercial banks have to face. For example, the inability to repay or the inability to repay in time after applying for a loan, malicious fraud of credit and other default phenomena will lead to bad debts in the bank and cause the bank's credit risk. Credit risk constitutes the focus and difficulty of bank risk management and is the core of bank risk management. How to predict credit risk becomes the key. Credit risk prediction is that commercial banks use scientific and reasonable analysis methods before issuing credit to comprehensively examine the current economic environment and economic situation of loan applicants, so as to comprehensively evaluate their repayment ability and repayment awareness, and then predict the credit risk of loan applicants according to certain standards.
[0003] The initial credit assessment is mainly based on the knowledge and experience accumulated by credit experts or professional credit personnel over the years. The main analysis process is to analyze the relevant information about the loan applicant collected in the consumer credit investigation to review the credit status of the loan subject, thereby evaluating its repayment ability and making a credit decision. The empirical method is a relatively traditional but efficient, convenient and effective assessment method for commercial banks to conduct credit assessments. However, the disadvantage of this method is that in the empirical credit assessment process, the professional experience and subjective judgment of experts are the decisive factors, lacking certain quantitative standards and subjectivity.
[0004] Therefore, scientific and effective credit risk prediction methods are used to improve the efficiency of credit assessment and enhance the credibility of credit assessment, which is an effective means of credit risk management. Building a mature and robust model to assist in assessment, while being able to effectively use multiple credit indicators as the basis for assessment, and still being able to reflect better prediction capabilities for smaller samples is a technical problem that technical personnel in this field urgently need to solve. Summary of the invention
[0005] To this end, the present invention proposes to establish a prediction model, which inputs the financial data of the loan subject, and the model outputs whether it will generate credit risk through analysis. The present invention adopts the support vector classifier method to train the sample data, and in order to optimize the parameters of the support vector classifier, the particle swarm optimization method is used to select the optimal parameter combination.
[0006] To achieve the above objectives, in a first aspect, the present invention provides a credit risk prediction method based on particle swarm optimization, comprising:
[0007] Obtaining indicator data in a credit management system, and dividing the indicator data into a training set and a test set;
[0008] The training set is used to train various parameters of the learning model; the test set is used to perform out-of-sample verification, and the prediction effect of the learning model on the loan default risk is judged based on the prediction effect of the test set;
[0009] Filter the index data according to the set rules, and substitute the filtered index data into the set function;
[0010] The setting function is iteratively calculated by a particle swarm intelligent algorithm according to the filtered index data, and the optimal solution is iteratively searched to obtain the optimal parameters of the setting function to generate a learning model;
[0011] The prediction effect of the learning model on the default risk is judged based on the prediction effect of the test set, and the target prediction model is obtained based on the judgment result.
[0012] Furthermore, the obtaining of credit-related data in the credit management system includes:
[0013] By random sampling, loan samples and samples with default risk are selected, and the loan samples and the samples with default risk are mixed in a set ratio, and the set ratio is 2 to 1.
[0014] Furthermore, the indicator data include: loan size, term, basic customer characteristic data, asset data, transaction behavior data, financial management, third-party custody data and other basic indicator data.
[0015] Furthermore, the setting function adopts a radial basis kernel function:
[0016]
[0017] Among them, xi is the input variable, xi∈R n , y represents the category variable of the classification result, yi∈{+1,-1}. The training data {xi, yi} composed of xi and yi, For a common sample point αi=0, the specific sample point xi corresponds to b * is the classification threshold.
[0018] Furthermore, the screening of the indicator data according to the set rules includes:
[0019] Feature screening is performed based on the characteristics of the indicator data, and the feature screening is performed based on a comprehensive analysis of static factors and dynamic factors of the customer related to credit risk; the static factors include the customer's education, occupation, age, education and other fixed factors; the dynamic behavioral factors include: based on various financial indicators and operational behaviors of the customer in commercial banks.
[0020] Furthermore, the specific process of generating the learning model includes:
[0021] The various indicators are idealized into particles, and the particles contain position information and speed information, and the position information and the speed information are represented by vectors; the particle parameter search range is set, the speed and position of the particle are initialized according to the set range, and the fitness function is called to calculate the fitness; the individual optimum and the global optimum are initialized according to the calculation results; the particle calculates a new position value and speed value according to the two values.
[0022] Furthermore, the particles calculate new position values and speed values according to the two values, including: the particles approach the optimal position of the group in each step of updating, and after several iterations, all particles are concentrated near the optimal value of the group; the particle swarm algorithm is an intelligent optimization algorithm proposed to simulate the foraging behavior of bird flocks, starting from a group of random solutions, exchanging information, and iteratively searching for the optimal solution; the particle swarm intelligent algorithm is selected to perform parameter optimization on the kernel function parameters and the penalty factor parameters, and the algorithm iteration termination conditions are determined according to the actual application to obtain the optimal parameters of the kernel function.
[0023] In a second aspect, the present invention also provides a credit risk prediction device based on particle swarm optimization, comprising: a data partitioning unit, a data screening unit, a computing unit, and an output unit;
[0024] The data division unit is used to obtain the index data in the credit management system and divide the index data into a training set and a test set;
[0025] The training set is used to train various parameters of the learning model; the test set is used to perform out-of-sample verification, and the prediction effect of the learning model on the loan default risk is judged based on the prediction effect of the test set;
[0026] The data screening unit is used to screen the index data according to the set rules and substitute the screened index data into the set function;
[0027] The computing unit is used to iteratively calculate the setting function according to the screened index data through a particle swarm intelligent algorithm, iteratively search for the optimal solution, and obtain the optimal parameters of the setting function to generate a learning model;
[0028] The output unit is used to judge the prediction effect of the learning model on the default risk based on the prediction effect of the test set, and obtain the target prediction model based on the judgment result.
[0029] In a third aspect, the present invention provides a computing device, comprising:
[0030] A memory for storing program instructions;
[0031] The processor is used to call the program instructions stored in the memory and execute the above-mentioned credit risk prediction method based on particle swarm optimization according to the obtained program instructions.
[0032] In a fourth aspect, the present invention provides a computer-readable storage medium, including computer-readable instructions. When a computer reads and executes the computer-readable instructions, the above-mentioned credit risk prediction method based on particle swarm optimization is implemented. The readable medium may be multiple, and the multiple readable media can operate independently of each other.
[0033] Compared with the existing technology, this solution innovatively uses the particle swarm optimization support vector machine method and applies it in the field of commercial bank credit risk prediction.
[0034] In terms of technology, compared with the manual evaluation method based on experience, a more intelligent support vector machine method is adopted to realize risk assessment. At the same time, it is proposed to use the particle swarm algorithm to optimize the support vector machine kernel function parameters. Compared with the traditional support vector machine, it has a better evaluation effect. The SVM model based on PSO optimization will have better performance in the evaluation model of personal credit risk prediction.
[0035] In terms of application, the present invention can obtain basic credit data of commercial bank customers and use scientific and effective prediction methods to more effectively predict credit risks and reduce bank bad debts caused by credit problems. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0037] Figure 1 A schematic diagram of a flow chart of an embodiment of a credit risk prediction method based on particle swarm optimization provided by the present invention;
[0038] Figure 2 An overall flow chart of an embodiment of a credit risk prediction method based on particle swarm optimization provided by the present invention;
[0039] Figure 3 A specific implementation flow chart of an embodiment of the credit risk prediction method based on particle swarm optimization provided by the present invention;
[0040] Figure 4 A structural block diagram of an embodiment of a credit risk prediction device based on particle swarm optimization provided by the present invention;
[0041] Figure 5 The present invention provides a schematic structural diagram of an electronic device. DETAILED DESCRIPTION
[0042] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be interpreted as limiting the present application.
[0043] It will be understood by those skilled in the art that, unless otherwise stated, the singular forms "a", "an" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present application refers to the presence of features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.
[0044] In order to make the objectives, technical solutions and advantages of the present application more clear, the specific implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.
[0045] The core of the present invention is to provide a credit risk prediction method based on particle swarm optimization. Figure 1 The flowchart of the credit risk prediction method based on particle swarm optimization provided by this embodiment is shown in FIG. Figure 1 The technical solution of this embodiment is described in detail:
[0046] Step S1: Obtain indicator data in the credit management system, and divide the indicator data into a training set and a test set;
[0047] The training set is used to train various parameters of the learning model; the test set is used to perform out-of-sample verification, and the prediction effect of the learning model on the loan default risk is judged based on the prediction effect of the test set;
[0048] Step S2: screening the indicator data according to the set rules, and substituting the screened indicator data into the set function;
[0049] Step S3: the setting function is iteratively calculated by a particle swarm intelligent algorithm according to the screened index data, the optimal solution is iteratively searched, and the optimal parameters of the setting function are obtained to generate a learning model;
[0050] Step S4: judging the prediction effect of the learning model on the default risk based on the prediction effect of the test set, and obtaining the target prediction model based on the judgment result.
[0051] In a preferred embodiment, the obtaining of credit-related data in the credit management system includes:
[0052] By random sampling, loan samples and samples with default risk are selected, and the loan samples and the samples with default risk are mixed in a set ratio, and the set ratio is 2 to 1.
[0053] In a preferred embodiment, the indicator data includes: loan size, term, basic characteristics of customers, asset data, transaction behavior data, financial management, third-party custody data and other basic indicator data. In a preferred embodiment, the setting function adopts a radial basis kernel function:
[0054]
[0055] Among them, xi is the input variable, xi∈R n , y represents the category variable of the classification result, yi∈{+1,-1}. The training data {xi, yi} composed of xi and yi, For a common sample point αi=0, the specific sample point xi corresponds to b * is the classification threshold.
[0056] In a preferred embodiment, the screening of the indicator data according to the set rules includes:
[0057] Feature screening is performed based on the characteristics of the indicator data, and the feature screening is performed based on a comprehensive analysis of static factors and dynamic factors of the customer related to credit risk; the static factors include the customer's education, occupation, age, education and other fixed factors; the dynamic behavioral factors include: based on various financial indicators and operational behaviors of the customer in commercial banks.
[0058] In a preferred embodiment, the specific process of generating the learning model includes: idealizing the various indicators into particles, the particles contain position information and speed information, and the position information and the speed information are represented by vectors; setting the particle parameter search range, initializing the speed and position of the particle according to the set range, and calling the fitness function to perform fitness calculation; initializing individual optimal and global optimal according to the calculation results; the particle calculates new position value and speed value based on the two values.
[0059] In a preferred embodiment, the particles calculate new position values and speed values according to the two values, including: the particles approach the optimal position of the group in each step of the update, and after several iterations, all particles are concentrated near the optimal value of the group; the particle swarm algorithm is an intelligent optimization algorithm proposed to simulate the foraging behavior of bird flocks, starting from a group of random solutions, exchanging information, and iteratively searching for the optimal solution; the particle swarm intelligent algorithm is selected to perform parameter optimization on the kernel function parameters and the penalty factor parameters, and the algorithm iteration termination conditions are determined according to the actual application to obtain the optimal parameters of the kernel function.
[0060] In the solution of this embodiment, the particle swarm optimization support vector machine method is innovatively used and applied in the field of commercial bank credit risk prediction. Compared with the manual evaluation method based on experience, a more intelligent support vector machine method is used to achieve risk assessment. At the same time, it is proposed to use the particle swarm algorithm to optimize the support vector machine kernel function parameters. Compared with the traditional support vector machine, it has a better evaluation effect. The SVM model based on PSO optimization will have better performance in the evaluation model of personal credit risk prediction. This embodiment can obtain the basic credit data of commercial bank customers and use scientific and effective prediction methods to more effectively predict credit risks and reduce bank bad debts caused by credit problems.
[0061] For ease of understanding, a specific example is provided below to further explain the present application.
[0062] Embodiment 1:
[0063] Figure 2 The figure shows the overall flow chart provided by this embodiment. Figure 3 For a specific implementation flow chart of this embodiment, see Figure 2 and Figure 3 include:
[0064] Step S101: Obtain credit-related data from the commercial bank's credit management system, core system, and bank credit system. Here, personal loan data is taken as an example, including the size and term of the loan, basic characteristic data of the customer, asset data, transaction behavior data, financial management, third-party custody data and other basic indicator data. A random sampling method is used to extract normal personal loan samples and samples with default risk, with a ratio of 2 to 1, and the samples are divided into a training set and a test set. The training set is mainly used to train the various parameters of the learning model, and then out-of-sample verification is performed on the test set. The prediction effect of the learning model on the personal loan default risk is judged based on the prediction effect of the test set.
[0065] Step S201: For many indicators of personal credit, since there is a correlation between the indicators, different indicators have different degrees of influence on the final result. Therefore, the relevant indicators are screened and used for modeling, so that the model has a better evaluation effect and can also take into account the evaluation efficiency. In feature screening, a comprehensive analysis should be conducted based on the static factors and dynamic factors of customers related to credit risk to achieve a better evaluation of the prediction effect. The static factors mainly include relatively fixed factors such as the customer's education, occupation, age, and education. The dynamic behavioral factors are mainly based on the various financial indicators and operational behaviors of the customer in the commercial bank. The dynamic behavioral factors are conducive to observing the customer's real cash flow, income and risk preference based on the customer's transaction behavior and investment behavior. For the application of personal credit of commercial banks, the features in Table 1 below can be selected as the input of the prediction evaluation.
[0066]
[0067]
[0068] Table 1
[0069] Step S301: Particle swarm optimization of SVM parameters is divided into the following two steps:
[0070] Step S311: determining the kernel function;
[0071] At present, the commonly used SVM kernel functions include linear kernel function, polynomial kernel function, two-layer neural network kernel function, and radial basis (Gaussian) kernel function. Since the radial basis kernel function has better learning ability in application and the value of the radial basis kernel function is easier to calculate, the radial basis kernel function is selected as the inner product function of the support vector machine.
[0072] When the support vector machine uses the radial basis kernel function, the optimal classification surface is:
[0073]
[0074] The specific decision-making process is: f(x)≥0.
[0075] Step S321: particle swarm intelligent algorithm optimization;
[0076] In the optimization iteration of the particle swarm intelligent algorithm, particles update their speed and position according to the following formula:
[0077] V i =ω×V i +c1×rand()×(pbest i -x i )+c2×rand()×(gbest i -x i )
[0078] x i =x i +V i
[0079] Where M is the number of particles in the group; V i is the current velocity of the particle; x i is the current position of the particle; pbest and gbest represent the optimal position of the individual and the optimal position of the population. ω, c1 and c2 are the parameters of the particle swarm algorithm. Set the PSO parameters as follows: C search range 1-1000, search range 0.00001-1, inertia factor 0.4, initial particle swarm size N 30, acceleration factor 1.7, acceleration factor 1.5, speed upper limit 1, speed lower limit -1, maximum number of iterations 100.
[0080] Step S401: Based on the prediction results of the test set, determine the prediction effect of the learning model on the personal loan default risk, analyze the model prediction accuracy, and obtain the final prediction model.
[0081] Next, the particle swarm intelligence algorithm involved in this application is explained to achieve a further understanding of the technical solution of this application:
[0082] The basic idea of the particle swarm intelligence algorithm is: by idealizing each individual in the bird flock as a particle, the movement of each particle can be extended to N-dimensional space. Each individual carries position information and speed information. The position can be represented by the vector X. i =(X1,X2,...,X N ), the velocity can be represented by the vector V i =(V1, V2, ..., V N ). Each particle can search in a certain space, which is mathematically explained as X iThe value of must be limited to a specific range. The fitness value of the particle can be obtained through a specific objective function. According to the fitness value, whether the position of the particle is the best position can be evaluated. Each particle records the best position it has found, recorded as pbest. Due to information sharing between particles, the particle also knows the best position that has been found in the entire population, recorded as gbest. Pbest and gbest determine the speed and direction of the particle's next step of movement. After finding pbest and gbest, the particle calculates the new position and speed values based on these two values. Repeating the above process, the particle gradually approaches the optimal position of the group in each step of the update. After multiple iterations, all particles are concentrated near the optimal value of the group. The particle swarm algorithm is an intelligent optimization algorithm proposed to simulate the foraging behavior of bird flocks. It starts from a set of random solutions, exchanges information, and iteratively searches for the optimal solution. The particle swarm intelligent algorithm is selected to optimize the kernel function parameters σ and the penalty factor c parameters. The algorithm iteration termination conditions are determined according to the actual application to obtain the optimal parameters of the SVM kernel function.
[0083] In one embodiment, the present invention provides a credit risk prediction device based on particle swarm optimization, see Figure 4 The figure shows a structural block diagram of a device for predicting credit risk based on particle swarm optimization provided by this embodiment, the device comprising: a data partitioning unit 10, a data screening unit 20, a computing unit 30, and an output unit 40;
[0084] The data division unit is used to obtain the index data in the credit management system and divide the index data into a training set and a test set;
[0085] The training set is used to train various parameters of the learning model; the test set is used to perform out-of-sample verification, and the prediction effect of the learning model on the loan default risk is judged based on the prediction effect of the test set;
[0086] The data screening unit is used to screen the index data according to the set rules and substitute the screened index data into the set function;
[0087] The computing unit is used to iteratively calculate the setting function according to the screened index data through a particle swarm intelligent algorithm, iteratively search for the optimal solution, and obtain the optimal parameters of the setting function to generate a learning model;
[0088] The output unit is used to judge the prediction effect of the learning model on the default risk based on the prediction effect of the test set, and obtain the target prediction model based on the judgment result.
[0089] In the solution of this embodiment, the particle swarm optimization support vector machine method is innovatively used and applied in the field of commercial bank credit risk prediction. Compared with the manual evaluation method based on experience, a more intelligent support vector machine method is used to achieve risk assessment. At the same time, it is proposed to use the particle swarm algorithm to optimize the support vector machine kernel function parameters. Compared with the traditional support vector machine, it has a better evaluation effect. The SVM model based on PSO optimization will have better performance in the evaluation model of personal credit risk prediction. This embodiment can obtain the basic credit data of commercial bank customers and use scientific and effective prediction methods to more effectively predict credit risks and reduce bank bad debts caused by credit problems.
[0090] The particle swarm optimization-based credit risk prediction device of this embodiment is used to implement the aforementioned particle swarm optimization-based credit risk prediction method. Therefore, the specific implementation method of the particle swarm optimization-based credit risk prediction device can be seen in the specific implementation method of the particle swarm optimization-based credit risk prediction method in the previous text. For example, the data partitioning unit 10 is used to implement step S1 in the above-mentioned particle swarm optimization-based credit risk prediction method, and the data screening unit 20 is used to implement step S2 in the above-mentioned particle swarm optimization-based credit risk prediction method, etc. Therefore, its specific implementation method can refer to the description of the corresponding various parts of the embodiment. The device provided by the present invention is attached to the method provided by the present invention, and its specific effects can refer to the above-mentioned method, which will not be repeated here.
[0091] In one embodiment, a computing device is provided, comprising:
[0092] A memory for storing program instructions;
[0093] A processor is used to implement the steps of any one of the above-mentioned credit risk prediction methods based on particle swarm optimization when executing the computer program.
[0094] As an example, Figure 5 A schematic diagram of the structure of an electronic device applicable to the embodiment of the present application is shown. Figure 5 As shown, the electronic device 2000 includes: a processor 2001 and a memory 2003. The processor 2001 and the memory 2003 are connected, such as through a bus 2002. Optionally, the electronic device 2000 may also include a transceiver 2004. It should be noted that in actual applications, the transceiver 2004 is not limited to one, and the structure of the electronic device 2000 does not constitute a limitation on the embodiments of the present application.
[0095] The processor 2001 is applied in the embodiment of the present application to implement the method shown in the above method embodiment. The transceiver 2004 may include a receiver and a transmitter. The transceiver 2004 is applied in the embodiment of the present application to implement the function of the electronic device of the embodiment of the present application communicating with other devices when executed.
[0096] Processor 2001 may be a CPU (Central Processing Unit), a general purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 2001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.
[0097] The bus 2002 may include a path to transmit information between the above components. The bus 2002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 2002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.
[0098] The memory 2003 can be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to these.
[0099] Optionally, the memory 2003 is used to store application code for executing the solution of the present application, and the execution is controlled by the processor 2001. The processor 2001 is used to execute the application code stored in the memory 2003 to implement the investment and financing project recommendation method provided in any embodiment of the present application.
[0100] The electronic device provided in the embodiment of the present application is applicable to any embodiment of the above method and will not be described in detail here.
[0101] The present invention provides a credit risk prediction method based on particle swarm optimization, which has obvious advantages and beneficial effects compared with the prior art. It innovatively uses the particle swarm optimization support vector machine method and applies it to the field of commercial bank credit risk prediction. Compared with the manual evaluation method based on experience, a more intelligent support vector machine method is used to achieve risk assessment. At the same time, it is proposed to use a particle swarm algorithm to optimize the support vector machine kernel function parameters. Compared with the traditional support vector machine, it has a better evaluation effect. The SVM model based on PSO optimization will have better performance in the evaluation model of personal credit risk prediction. This embodiment can obtain the basic credit data of commercial bank customers and use scientific and effective prediction methods to more effectively predict credit risks and reduce bank bad debts caused by credit problems.
[0102] In one embodiment, a computer-readable storage medium is provided, including computer-readable instructions. When a computer reads and executes the computer-readable instructions, the steps of the above-mentioned credit risk prediction method based on particle swarm optimization are implemented. The credit risk prediction method based on particle swarm optimization provided by the present invention has obvious advantages and beneficial effects compared with the prior art. It innovatively uses the particle swarm optimization support vector machine method and applies it in the field of commercial bank credit risk prediction. Compared with the manual evaluation method based on experience, a more intelligent support vector machine method is used to achieve risk assessment. At the same time, it is proposed to use the particle swarm algorithm to optimize the support vector machine kernel function parameters. Compared with the traditional support vector machine, it has a better evaluation effect. The SVM model based on PSO optimization will have better performance in the evaluation model of personal credit risk prediction. This embodiment can obtain the basic credit data of commercial bank customers and use scientific and effective prediction methods to more effectively predict credit risks and reduce bank bad debts caused by credit problems.
[0103] Through the description of the above implementation methods, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the specific device can be divided into different functional modules to complete all or part of the functions described above.
[0104] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0105] In several embodiments covered by the present application, it should be understood that the disclosed devices, equipment and methods can be implemented in other ways. For example, the division of modules is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, or it can be an electrical, mechanical or other form of connection.
[0106] The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of the present application.
[0107] In addition, each functional module in each embodiment of the present application can be implemented in the form of hardware or in the form of software functional modules. If these functional modules are implemented in the form of software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application is essentially or part of the prior art that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a computer program product, which includes one or more computer instructions. When loading and executing computer program instructions on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer program product is stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, a computer instruction can be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or a data center that includes one or more available media integrations. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium (eg, a solid state disk (SSD)).
[0108] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.
Claims
1. A credit risk prediction method based on particle swarm optimization, characterized in that: include: Obtaining indicator data in a credit management system, and dividing the indicator data into a training set and a test set; The training set is used to train various parameters of the learning model; the test set is used to perform out-of-sample verification, and the prediction effect of the learning model on the loan default risk is judged based on the prediction effect of the test set; Filter the index data according to the set rules, and substitute the filtered index data into the set function; The setting function is iteratively calculated by a particle swarm intelligent algorithm according to the filtered index data, and the optimal solution is iteratively searched to obtain the optimal parameters of the setting function to generate a learning model; The prediction effect of the learning model on the default risk is judged based on the prediction effect of the test set, and the target prediction model is obtained based on the judgment result.
2. The credit risk prediction method based on particle swarm optimization according to claim 1, characterized in that: The obtaining of credit-related data in the credit management system includes: By random sampling, loan samples and samples with default risk are selected, and the loan samples and the samples with default risk are mixed in a set ratio, and the set ratio is 2 to 1.
3. The credit risk prediction method based on particle swarm optimization according to claim 2, characterized in that: The indicator data include: loan size, term, basic customer characteristic data, asset data, transaction behavior data, financial management, third-party custody data and other basic indicator data.
4. The credit risk prediction method based on particle swarm optimization according to claim 2, characterized in that: The setting function adopts the radial basis kernel function as follows: Among them, xi is the input variable, xi∈R n , y represents the categorical variable of the classification result, yi∈{+1,-1}. The training data {xi, yi} composed of xi and yi, For a common sample point αi=0, the specific sample point xi corresponds to b * is the classification threshold.
5. The credit risk prediction method based on particle swarm optimization according to any one of claims 1 and 2, characterized in that: The screening of the indicator data according to the set rules includes: Feature screening is performed based on the characteristics of the indicator data, and the feature screening is performed based on a comprehensive analysis of static factors and dynamic factors of the customer related to credit risk; the static factors include the customer's education, occupation, age, education and other fixed factors; the dynamic behavioral factors include: based on various financial indicators and operational behaviors of the customer in commercial banks.
6. The credit risk prediction method based on particle swarm optimization according to claim 1, characterized in that: The specific process of generating the learning model includes: The various indicators are idealized into particles, and the particles contain position information and speed information, and the position information and the speed information are represented by vectors; the particle parameter search range is set, the speed and position of the particle are initialized according to the set range, and the fitness function is called to calculate the fitness; the individual optimum and the global optimum are initialized according to the calculation results; the particle calculates a new position value and speed value according to the two values.
7. The credit risk prediction method based on particle swarm optimization according to claim 1, characterized in that: The particles calculate new position values and speed values according to the two values, including: the particles approach the optimal position of the group in each step of updating, and after several iterations, all particles are concentrated near the optimal value of the group; the particle swarm algorithm is an intelligent optimization algorithm proposed to simulate the foraging behavior of bird flocks, starting from a group of random solutions, exchanging information, and iteratively searching for the optimal solution; the particle swarm intelligent algorithm is selected to perform parameter optimization on the kernel function parameters and the penalty factor parameters, and the algorithm iteration termination conditions are determined according to the actual application to obtain the optimal parameters of the kernel function.
8. A credit risk prediction device based on particle swarm optimization, characterized in that: The device comprises: a data division unit, a data screening unit, a calculation unit, and an output unit; The data division unit is used to obtain the index data in the credit management system and divide the index data into a training set and a test set; The training set is used to train various parameters of the learning model; the test set is used to perform out-of-sample verification, and the prediction effect of the learning model on the loan default risk is judged based on the prediction effect of the test set; The data screening unit is used to screen the index data according to the set rules and substitute the screened index data into the set function; The computing unit is used to iteratively calculate the setting function according to the screened index data through a particle swarm intelligent algorithm, iteratively search for the optimal solution, and obtain the optimal parameters of the setting function to generate a learning model; The output unit is used to judge the prediction effect of the learning model on the default risk based on the prediction effect of the test set, and obtain the target prediction model based on the judgment result.
9. A computing device, characterized in that include: A memory for storing program instructions; A processor is used to call the program instructions stored in the memory, and execute the method according to any one of claims 1 to 7 according to the obtained program instructions.
10. A computer-readable storage medium, characterized in that: It includes computer-readable instructions. When a computer reads and executes the computer-readable instructions, it implements the method as claimed in any one of claims 1 to 7. There can be multiple readable media, and the multiple readable media can operate independently of each other.
Citation Information
Cited By
Performance model parameter simulation optimization method, electronic equipment and storage medium
CN121094654A
