User financial risk identification model generation method, device and electronic equipment

Abstract

The present disclosure relates to a user financial risk identification model generation method and device, electronic equipment and computer readable medium. The method comprises: obtaining financial data of a plurality of historical users, wherein the financial data comprises a default state; selecting a plurality of positive sample users from the plurality of historical users according to the default state; generating a hypersphere equation set through the financial data of the plurality of positive sample users and an initial hypersphere equation; solving the hypersphere equation to obtain an optimal solution; and generating a user financial risk identification model based on the optimal solution, wherein the user financial risk identification model is used to determine the financial default risk probability of a user. The user financial risk identification model generation method and device, electronic equipment and computer readable medium disclosed by the present disclosure can adapt to the machine learning model training process when the number of samples is unbalanced, and can also generate an accurate and efficient user financial risk identification model using unbalanced sample data.

Description

Technical Field

[0001] The present disclosure relates to the field of computer information processing, and in particular, to a method, device, electronic device, and computer-readable medium for generating a user financial risk identification model. Background Art

[0002] In the typical application of machine learning models, the user first identifies a machine learning model for a specific category or algorithm. Based on the specific problem they want to solve, they input specific data, and the machine learning model is assigned a specific task. The model is then trained using this data, resulting in a machine learning model suitable for the specific task. Often, even if the same machine learning model algorithm is used, training it with different data will yield completely different models. Machine learning models typically learn from both positive and negative samples. Positive samples are samples corresponding to correctly classified categories, while negative samples can, in principle, be any sample that is not in the correct category. However, for models in the financial sector, particularly those assessing user behavior and financial risk, the number of positive samples often far outnumbers the number of negative samples (for example, 1:99), creating a significant imbalance between positive and negative samples.

[0003] In existing technology, when training machine learning models, this situation is often addressed through data processing methods such as downsampling or oversampling to manually adjust the ratio of positive and negative samples. However, downsampling can lead to information loss, while oversampling can add useless information. Furthermore, if training is performed directly with the original sample ratio, the binary classification model will treat the smaller number of positive samples as noise, resulting in a poorly performing model.

[0004] Therefore, a new method, device, electronic device and computer-readable medium for generating a user financial risk identification model are needed.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0006] In view of this, the present disclosure provides a method, device, electronic device and computer-readable medium for generating a user financial risk identification model, which can adapt to the machine learning model training process when the sample size is unbalanced, and can also use unbalanced sample data to generate an accurate and efficient user financial risk identification model.

[0007] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by practice of the present disclosure.

[0008] According to one aspect of the present disclosure, a method for generating a user financial risk identification model is proposed, the method comprising: obtaining financial data of a plurality of historical users, the financial data including default status; screening out a plurality of positive sample users from the plurality of historical users based on the default status; generating a set of hypersphere equations using the financial data of the plurality of positive sample users and an initial hypersphere equation; solving the hypersphere equation to obtain its optimal solution; and generating a user financial risk identification model based on the optimal solution, the user financial risk identification model being used to determine the user's financial default risk probability.

[0009] Optionally, generating a group of hypersphere equations using the financial data of the multiple positive sample users and the initial hypersphere equation includes: generating multi-dimensional financial data feature values ​​of the multiple positive sample users using the financial data of the multiple positive sample users and a preset multi-dimensional feature vector baseline value; and generating the group of hypersphere equations based on the multi-dimensional financial data feature values ​​of the multiple positive sample users and the initial hypersphere equation.

[0010] Optionally, generating the multi-dimensional financial data characteristic values ​​of the multiple positive sample users through the financial data of the multiple positive sample users and the preset multi-dimensional feature vector benchmark values ​​includes: classifying the financial data of the multiple positive sample users according to preset dimensions to generate multiple multi-dimensional financial data; and comparing the multiple multi-dimensional financial data with the preset multi-dimensional feature vector benchmark values ​​to generate the multi-dimensional financial data characteristic values ​​of the multiple positive sample users.

[0011] Optionally, the method further includes: determining a characteristic dimension of the financial data of the historical user; and determining a reference value of the multi-dimensional characteristic vector based on the characteristic dimension and numerical values ​​of the financial data of multiple historical users corresponding to the dimension.

[0012] Optionally, generating the set of hypersphere equations based on the multi-dimensional financial data feature values ​​of the multiple positive sample users and the initial hypersphere equation includes: determining the corresponding multiple position coordinates in a multi-dimensional coordinate system based on the multi-dimensional financial data feature values ​​of the multiple positive sample users; and substituting the multiple position coordinates into the initial hypersphere equation respectively to generate the set of hypersphere equations.

[0013] Optionally, obtaining the optimal solution of the hypersphere equation includes: determining a target slack variable threshold and an optimization target; and solving the initial hypersphere equation through an optimization algorithm based on the target slack variable threshold and the optimization target to obtain the optimal solution of the initial hypersphere equation; wherein the optimization target is that all positive sample users are within the hypersphere.

[0014] Optionally, generating the user financial risk identification model based on the optimal solution includes: generating a target hypersphere equation based on the optimal solution; and generating the user financial risk identification model based on the target hypersphere equation.

[0015] Optionally, the method further includes: inputting the financial data of the current user into the target hypersphere equation of the financial risk identification model; obtaining the distance between the current user and the center of the target hypersphere equation; and determining the financial default risk probability of the user based on the center distance.

[0016] Optionally, obtaining the distance between the current user and the center of the target hypersphere equation includes: solving the target hypersphere equation using a Lagrange dual method to obtain the distance between the user and the center of the hypersphere.

[0017] Optionally, determining the user's financial default risk probability based on the sphere center distance includes: when the sphere center distance is greater than the target hypersphere radius, determining the user as a low-risk default user; and when the sphere center distance is less than or equal to the target hypersphere radius, determining the user as a high-risk default user.

[0018] According to one aspect of the present disclosure, a device for generating a user financial risk identification model is proposed, which includes: a data module for obtaining financial data of multiple historical users, wherein the financial data includes default status; a screening module for screening out multiple positive sample users from the multiple historical users according to the default status; an equation module for generating a hypersphere equation group through the financial data of the multiple positive sample users and an initial hypersphere equation; a solution module for solving the hypersphere equation to obtain its optimal solution; and a model module for generating a user financial risk identification model based on the optimal solution, wherein the user financial risk identification model is used to determine the user's financial default risk probability.

[0019] Optionally, the equation module includes: a feature unit, used to generate multi-dimensional financial data feature values ​​of the multiple positive sample users through the financial data of the multiple positive sample users and a preset multi-dimensional feature vector baseline value; an equation group unit, used to generate the hypersphere equation group based on the multi-dimensional financial data feature values ​​of the multiple positive sample users and the initial hypersphere equation.

[0020] Optionally, the feature unit includes: a classification subunit, used to classify the financial data of the multiple positive sample users according to preset dimensions to generate multiple multi-dimensional financial data; and a comparison subunit, used to compare the multiple multi-dimensional financial data with a preset multi-dimensional feature vector benchmark value to generate multi-dimensional financial data feature values ​​of the multiple positive sample users.

[0021] Optionally, the feature unit further includes: a dimension subunit, used to determine the feature dimension of the financial data of the historical user; and a benchmark subunit, used to determine the multi-dimensional feature vector benchmark value based on the feature dimension and the numerical value of the financial data of multiple historical users corresponding to the dimension.

[0022] Optionally, the equation group unit includes: a coordinate sub-unit, used to determine the corresponding multiple position coordinates in the multi-dimensional coordinate system based on the multi-dimensional financial data feature values ​​of the multiple positive sample users; and a substitution sub-unit, used to substitute the multiple position coordinates into the initial hypersphere equation respectively to generate the hypersphere equation group.

[0023] Optionally, the solution module includes: a parameter unit, used to determine a target slack variable threshold and an optimization target; and a calculation unit, used to solve the initial hypersphere equation through an optimization algorithm based on the target slack variable threshold and the optimization target to obtain the optimal solution of the initial hypersphere equation; wherein the optimization target is that all positive sample users are within the hypersphere.

[0024] Optionally, the model module includes: a target unit, used to generate a target hypersphere equation based on the optimal solution; and a model unit, used to generate the user financial risk identification model based on the target hypersphere equation.

[0025] Optionally, it also includes: an input module for inputting the current user's financial data into the target hypersphere equation of the financial risk identification model; a distance module for obtaining the distance between the current user and the center of the target hypersphere equation; and a risk module for determining the user's financial default risk probability based on the center distance.

[0026] Optionally, the input module is further configured to solve the target hypersphere equation using a Lagrange dual method to obtain the user's sphere center distance.

[0027] Optionally, the risk module is further configured to determine that the user is a low-risk default user when the sphere center distance is greater than the target hypersphere radius; and to determine that the user is a high-risk default user when the sphere center distance is less than or equal to the target hypersphere radius.

[0028] According to one aspect of the present disclosure, an electronic device is proposed, which includes: one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described above.

[0029] According to one aspect of the present disclosure, a computer-readable medium is provided, on which a computer program is stored. When the program is executed by a processor, the method described above is implemented.

[0030] According to the user financial risk identification model generation method, device, electronic device and computer-readable medium disclosed in the present invention, multiple positive sample users are screened out from the multiple historical users according to the default status; a hypersphere equation group is generated through the financial data of the multiple positive sample users and the initial hypersphere equation; the hypersphere equation is calculated to obtain its optimal solution; and a user financial risk identification model is generated based on the optimal solution. The user financial risk identification model is used to determine the user's financial default risk probability in a manner that can adapt to the machine learning model training process when the sample size is unbalanced, and can also use unbalanced sample data to generate an accurate and efficient user financial risk identification model.

[0031] It should be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The above and other objects, features and advantages of the present disclosure will become more apparent by describing in detail exemplary embodiments thereof with reference to the accompanying drawings. The drawings described below are only some embodiments of the present disclosure, and it is obvious to those skilled in the art that other drawings can be derived from these drawings without inventive effort.

[0033] Figure 1 The present invention is a flowchart of a method for generating a user financial risk identification model according to an exemplary embodiment.

[0034] Figure 2 The figure is a flowchart of a method for generating a user financial risk identification model according to another exemplary embodiment.

[0035] Figure 3 is a schematic diagram of a method for generating a user financial risk identification model according to another exemplary embodiment.

[0036] Figure 4 The figure is a flowchart of a method for generating a user financial risk identification model according to another exemplary embodiment.

[0037] Figure 5 The present invention is a block diagram of a device for generating a user financial risk identification model according to an exemplary embodiment.

[0038] Figure 6 The present invention is a block diagram of a device for generating a user financial risk identification model according to an exemplary embodiment.

[0039] Figure 7 It is a block diagram of an electronic device according to an exemplary embodiment.

[0040] Figure 8 It is a block diagram of a computer-readable medium according to an exemplary embodiment. DETAILED DESCRIPTION

[0041] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be embodied in many forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art. Like reference numerals in the drawings represent like or similar parts, and thus repetitive description thereof will be omitted.

[0042] In addition, the described features, structures or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure can be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, well-known methods, devices, implementations or operations are not shown or described in detail to avoid blurring various aspects of the present disclosure.

[0043] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0044] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0045] It should be understood that although the terms first, second, third, etc. may be used herein to describe various components, these components should not be limited by these terms. These terms are used to distinguish one component from another. Thus, the first component discussed below can be referred to as the second component without departing from the teachings of the present disclosure. As used herein, the term "and / or" includes any one of the associated listed items and all combinations of one or more of them.

[0046] Those skilled in the art will understand that the drawings are merely schematic diagrams of example embodiments, and the modules or processes in the drawings are not necessarily necessary for implementing the present disclosure, and therefore cannot be used to limit the scope of protection of the present disclosure.

[0047] Figure 1FIG1 is a flow chart showing a method for generating a user financial risk identification model according to an exemplary embodiment. The method 10 for generating a user financial risk identification model includes at least steps S102 to S108.

[0048] like Figure 1 As shown, in S102, financial data of multiple historical users is obtained, wherein the financial data includes default status. The historical user financial data may be user financial data of a financial services company. The default status may indicate whether the user has defaulted on their outstanding debts. The default status may include defaulting users and users who are in default.

[0049] The default status can also indicate the user's other financial credit defaults, such as failure to pay certain fees on time, failure to process certain fines on time, etc.

[0050] In S104, a plurality of positive sample users are screened out from the plurality of historical users based on the default status, wherein the positive sample set may include financial data of users with a history of default.

[0051] In S106, a system of hypersphere equations is generated using the financial data of the multiple positive sample users and the initial hypersphere equation. Common outlier detection methods currently used in the industry to address the imbalance between positive and negative samples are either statistically based or unsupervised. Statistical methods are generally ineffective, while unsupervised methods do not fully utilize the positive sample label information. The OneClassSVM algorithm, as a supervised outlier detection algorithm, is more suitable for the above scenario than traditional outlier processing methods.

[0052] In one embodiment, generating a set of hypersphere equations using the financial data of the multiple positive sample users and the initial hypersphere equation includes: generating multi-dimensional financial data feature values ​​of the multiple positive sample users using the financial data of the multiple positive sample users and a preset multi-dimensional feature vector baseline value; and generating the set of hypersphere equations based on the multi-dimensional financial data feature values ​​of the multiple positive sample users and the initial hypersphere equation.

[0053] A hypersphere, also known as a hypersphere or n-dimensional sphere, is a generalization of a regular sphere to any dimension. It is an n-dimensional manifold in (n+1)-dimensional space. Specifically, a 0-dimensional sphere is two points on a straight line, a 1-dimensional sphere is a circle on a plane, and a 2-dimensional sphere is a regular sphere in three-dimensional space. Spheres with dimensions higher than 2 are called hyperspheres.

[0054] More specifically, the parameters of the hypersphere can be defined as follows: an n-dimensional sphere with its center at the origin and a unit radius is called a unit n-dimensional sphere, denoted by Sn. Symbolically, it is:

[0055] Sn={x∈IRn+1:||x||=1};

[0056] A hypersphere is the surface or boundary of an n-dimensional sphere, a type of n-dimensional manifold.

[0057] For any natural number n, an n-dimensional sphere of radius r is defined as the set of all points in (n+1)-dimensional Euclidean space whose distance from a fixed point is equal to a constant r, where r can be any positive real number. It is an n-dimensional manifold in (n+1)-dimensional space.

[0058] The details of “generating a set of hypersphere equations using the financial data of the plurality of positive sample users and the initial hypersphere equation” will be discussed in detail. Figure 2 The details are described in the corresponding embodiments.

[0059] In S108, the hypersphere equation is obtained to obtain its optimal solution. This includes: determining a target slack variable threshold and an optimization target; and solving the initial hypersphere equation using an optimization algorithm based on the target slack variable threshold and the optimization target to obtain the optimal solution of the initial hypersphere equation. The optimization target is that all positive sample users are within the hypersphere.

[0060] The introduction of slack variables is often to facilitate solving in a larger feasible region. If it is 0, the problem converges to the original state; if it is greater than zero, the constraint is relaxed.

[0061] The study of linear programming problems is based on the standard form. Therefore, for a given non-standard linear programming problem, the mathematical model needs to be transformed into the standard form. Generally speaking, for different forms of linear programming models, some methods can be used to transform them into the standard form. Among them,

[0062] When a linear programming problem has "≤" ("≥") constraints, you can convert the inequality into an equation by adding (or subtracting) a new nonnegative variable to the left side of the inequality. This new nonnegative variable is called a slack variable (or residual variable), or generally referred to as a slack variable. In the objective function, the coefficient of the new slack variable is generally assumed to be zero.

[0063] In S110 , a user financial risk identification model is generated based on the optimal solution, and the user financial risk identification model is used to determine the user's financial default risk probability. This includes: generating a target hypersphere equation based on the optimal solution; and generating the user financial risk identification model based on the target hypersphere equation.

[0064] As described above, the hypersphere formula can be used to construct an initial hypersphere equation based on the hypersphere equation and the multi-dimensional financial data feature values ​​of multiple users in the positive sample set. An optimization objective is determined, and the hypersphere method is continuously solved using thresholds on the slack variables until an optimal solution that meets the optimization objective is obtained. The target hypersphere equation is then generated using the parameters of the hypersphere equation corresponding to the optimal solution.

[0065] According to the user financial risk identification model generation method disclosed in the present invention, multiple positive sample users are screened out from the multiple historical users according to the default status; a hypersphere equation group is generated through the financial data of the multiple positive sample users and the initial hypersphere equation; the hypersphere equation is calculated to obtain its optimal solution; and a user financial risk identification model is generated based on the optimal solution. The user financial risk identification model is used to determine the user's financial default risk probability in a manner that can adapt to the machine learning model training process when the sample size is unbalanced, and can also use unbalanced sample data to generate an accurate and efficient user financial risk identification model.

[0066] It should be clearly understood that the present disclosure describes how to form and use specific examples, but the principles of the present disclosure are not limited to any details of these examples. On the contrary, based on the teachings of the contents disclosed in this disclosure, these principles can be applied to many other embodiments.

[0067] Figure 2 The figure is a flowchart of a method for generating a user financial risk identification model according to another exemplary embodiment. Figure 2 The process shown is Figure 1 Detailed description of S106 "generating a set of hypersphere equations by using the financial data of the plurality of positive sample users and the initial hypersphere equation" in the process shown.

[0068] like Figure 2 As shown, in S202, the financial data of the plurality of positive sample users are classified according to preset dimensions to generate a plurality of multi-dimensional financial data, wherein, for example, the financial characteristics of the users, such as income, age, gender, job category, etc., can be respectively used as a characteristic dimension.

[0069] In S204 , the plurality of multi-dimensional financial data are compared with a preset multi-dimensional feature vector reference value to generate multi-dimensional financial data feature values ​​of the plurality of positive sample users.

[0070] The method further includes: determining a characteristic dimension of the financial data of the historical user; and determining a reference value of the multi-dimensional characteristic vector based on the characteristic dimension and the numerical values ​​of the financial data of multiple historical users corresponding to the dimension.

[0071] More specifically, the average user income of the positive samples can be compared, and the characteristic value of the financial data in this dimension can be determined through values ​​such as variance or standard deviation.

[0072] In one embodiment, features such as age can be represented using a one-hot method and divided into segments of 0-10, 10-20..., >100. A representative value can be determined for each segment, for example, the representative value of 0-10 can be 1, the representative value of 10-20 can be 2, and so on.

[0073] In one embodiment, for features such as income, an average method may be used, for example, by using user financial data in a large number of positive sample sets to determine the average income of users in the positive sample set.

[0074] For example, the age of the user in the positive sample can be compared with the standard age range to determine the feature value of the age dimension.

[0075] In S206, a plurality of position coordinates corresponding to the plurality of positive sample users are determined in a multidimensional coordinate system based on the multidimensional financial data feature values ​​thereof. The multidimensional financial data feature values ​​are projected in the multidimensional coordinate system to determine their positions.

[0076] In S208 , the plurality of position coordinates are substituted into the initial hypersphere equations respectively to generate the hypersphere equation group.

[0077] Figure 3 is a schematic diagram of a method for generating a user financial risk identification model according to another exemplary embodiment.

[0078] like Figure 3 As shown, to put it simply, it can be understood as finding a hyperplane to circle the positive examples in the sample (target hypersphere). Prediction is to use this hyperplane to make decisions, and the samples in the target hypersphere are considered to be positive samples.

[0079] Assume that the parameters of the generated hypersphere are the center o and the corresponding hypersphere radius r>0, the hypersphere volume V(r) is minimized, and the center o is a linear combination of the support rows; it can be required that the data points x representing the user characteristics in all unclassified sample sets i The distance to the center is strictly less than r. At the same time, a slack variable ζ with a penalty coefficient C is constructed i , the optimization problem is shown as follows:

[0080] MinV(r)+C∑ζ i；

[0081] ||x i -o||≤r+ζ i , i=1,2,3,…m;

[0082] ζ i ≥0, i=1,2,3,…m.

[0083] After solving the Lagrange duality, find the distance from z to the center.

[0084] Basic Principle: The OneClassSVM method searches for an optimal hypersphere when training on positive sample data. This hypersphere can enclose as many positive samples as possible. Samples within the hypersphere are positive, while samples outside the hypersphere are negative. In practical applications, samples falling within this hypersphere are positive, while samples falling outside the hypersphere are negative.

[0085] Figure 4 The figure is a flowchart of a method for generating a user financial risk identification model according to another exemplary embodiment. Figure 4 The process shown is a further description of the method for generating a user financial risk identification model.

[0086] like Figure 4 As shown, in S402, the current user's financial data is input into the target hypersphere equation of the financial risk identification model. The user's financial data may first be subjected to multi-dimensional data processing. More specifically, the current user's financial data may be processed with reference to preset user feature dimensions to generate multi-dimensional feature values ​​for the current user.

[0087] In S404, the distance between the current user and the center of the target hypersphere equation is obtained. The target hypersphere equation can be solved using the Lagrange dual method to obtain the user's center distance. The Lagrange dual method is a method for finding the extreme value of a multivariate function whose variables are subject to one or more conditions. This method transforms an optimization problem with n variables and k constraints into an extreme value problem for a system of equations with n + k variables, where the variables are not constrained. This method introduces a new scalar unknown, the Lagrange multiplier: the coefficient of each vector in the linear combination of the gradients of the constraint equations. The proof of this method involves partial differentiation, total differentiation, or the chain method to find the value of the unknown that makes the differential of the assumed implicit function zero.

[0088] More specifically, we need to find the distance between the user's multi-dimensional feature value f(x, y, z) and the center of the hypersphere φ(x, y, z) = 0. The specific process is as follows:

[0089] 1. Make the Lagrangian function L = f (x, y, z) + λφ (x, y, z), where λ is called the Lagrange multiplier;

[0090] 2. Find the partial derivatives of L with respect to x, y, z, and λ, respectively, to obtain the system of equations and find the stationary point P(x, y, z);

[0091] If the maximum or minimum value of this practical problem exists, there is generally only one stationary point, so the maximum value can be found.

[0092] The distance between the sphere center of the user's multi-dimensional feature value f(x, y, z) is the minimum value of the above formula, which can be quickly obtained through Lagrangian.

[0093] In S406 , when the sphere center distance is greater than the radius of the target hypersphere, the user is determined to be a low-risk default user.

[0094] In S408, when the sphere center distance is less than or equal to the radius of the target hypersphere, the user is determined to be a high-risk default user.

[0095] Those skilled in the art will appreciate that all or part of the steps implementing the above embodiments can be implemented as a computer program executed by a CPU. When executed by the CPU, this computer program performs the functions defined in the above methods provided herein. The program can be stored in a computer-readable storage medium, such as a read-only memory, a magnetic disk, or an optical disk.

[0096] Furthermore, it should be noted that the aforementioned figures are merely illustrative of the processes included in the methods according to exemplary embodiments of the present disclosure and are not intended to be limiting. It is readily understood that the processes illustrated in the aforementioned figures do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0097] The following are embodiments of the apparatus disclosed herein, which can be used to implement the method embodiments disclosed herein. For details not disclosed in the apparatus embodiments disclosed herein, please refer to the method embodiments disclosed herein.

[0098] Figure 5 FIG. 1 is a block diagram of a device for generating a user financial risk identification model according to an exemplary embodiment. Figure 5 As shown, the user financial risk identification model generating device 50 includes: a data module 502 , a screening module 504 , an equation module 506 , a solution module 508 , and a model module 510 .

[0099] The data module 502 is used to obtain financial data of multiple historical users, wherein the financial data includes default status;

[0100] The screening module 504 is configured to screen out a plurality of positive sample users from the plurality of historical users according to the default status;

[0101] The equation module 506 is configured to generate a hypersphere equation group using the financial data of the plurality of positive sample users and the initial hypersphere equation;

[0102] The equation module 506 includes: a feature unit, which is used to generate multi-dimensional financial data feature values ​​of the multiple positive sample users through the financial data of the multiple positive sample users and a preset multi-dimensional feature vector benchmark value; the feature unit includes: a classification subunit, which is used to classify the financial data of the multiple positive sample users according to preset dimensions to generate multiple multi-dimensional financial data; and a comparison subunit, which is used to compare the multiple multi-dimensional financial data with the preset multi-dimensional feature vector benchmark value to generate the multi-dimensional financial data feature values ​​of the multiple positive sample users.

[0103] The feature unit further includes: a dimension subunit for determining a feature dimension of the historical user's financial data; and a benchmark subunit for determining a benchmark value of the multi-dimensional feature vector based on the feature dimension and the numerical value of the financial data of multiple historical users corresponding to the dimension. An equation system unit for generating the hypersphere equation system based on the feature values ​​of the multi-dimensional financial data of the multiple positive sample users and the initial hypersphere equation.

[0104] The equation group unit includes: a coordinate sub-unit, used to determine the corresponding multiple position coordinates in a multi-dimensional coordinate system based on the multi-dimensional financial data feature values ​​of the multiple positive sample users; and a substitution sub-unit, used to substitute the multiple position coordinates into the initial hypersphere equation respectively to generate the hypersphere equation group.

[0105] The solution module 508 is used to obtain the optimal solution of the hypersphere equation; the solution module 508 includes: a parameter unit, used to determine the target slack variable threshold and the optimization target; and a calculation unit, used to solve the initial hypersphere equation through an optimization algorithm based on the target slack variable threshold and the optimization target to obtain the optimal solution of the initial hypersphere equation; wherein the optimization target is that all positive sample users are within the hypersphere.

[0106] The model module 510 is used to generate a user financial risk identification model based on the optimal solution, and the user financial risk identification model is used to determine the user's financial default risk probability.

[0107] The model module 510 includes: a target unit, configured to generate a target hypersphere equation based on the optimal solution; and a model unit, configured to generate the user financial risk identification model based on the target hypersphere equation.

[0108] Figure 6 FIG. 1 is a block diagram of a device for generating a user financial risk identification model according to another exemplary embodiment. Figure 6 As shown, the user financial risk identification model generating device 60 further includes: an input module 602 , a distance module 604 , and a risk module 606 in the user financial risk identification model generating device 50 .

[0109] The input module 602 is used to input the current user's financial data into the target hypersphere equation of the financial risk identification model; the input module 602 is also used to solve the target hypersphere equation using the Lagrangian dual method to obtain the user's sphere center distance.

[0110] The distance module 604 is used to obtain the distance between the current user and the center of the target hypersphere equation; and

[0111] The risk module 606 is configured to determine the user's financial default risk probability based on the sphere center distance. The risk module 606 is further configured to determine the user as a low-risk default user when the sphere center distance is greater than the target hypersphere radius; and to determine the user as a high-risk default user when the sphere center distance is less than or equal to the target hypersphere radius.

[0112] According to the user financial risk identification model generation device disclosed in the present invention, multiple positive sample users are screened out from the multiple historical users according to the default status; a hypersphere equation group is generated through the financial data of the multiple positive sample users and the initial hypersphere equation; the hypersphere equation is calculated to obtain its optimal solution; and a user financial risk identification model is generated based on the optimal solution. The user financial risk identification model is used to determine the user's financial default risk probability in a manner that can adapt to the machine learning model training process when the sample size is unbalanced, and can also use unbalanced sample data to generate an accurate and efficient user financial risk identification model.

[0113] Figure 7 It is a block diagram of an electronic device according to an exemplary embodiment.

[0114] Refer to the following Figure 7 700 according to this embodiment of the present disclosure will be described. Figure 7 The electronic device 700 shown is merely an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0115] like Figure 7 As shown, electronic device 700 is implemented as a general-purpose computing device. Components of electronic device 700 may include, but are not limited to, at least one processing unit 710, at least one storage unit 720, a bus 730 connecting various system components (including storage unit 720 and processing unit 710), and a display unit 740.

[0116] The storage unit stores program codes, which can be executed by the processing unit 710, so that the processing unit 710 executes the steps according to various exemplary embodiments of the present disclosure described in the electronic prescription circulation processing method section of this specification. For example, the processing unit 710 can execute the following steps: Figure 1 , Figure 2 , Figure 4 Follow the steps shown in .

[0117] The storage unit 720 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 7201 and / or a cache memory unit 7202 , and may further include a read-only memory unit (ROM) 7203 .

[0118] The storage unit 720 may also include a program / utility 7204 having a set (at least one) of program modules 7205, such program modules 7205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each of which or some combination may include an implementation of a network environment.

[0119] Bus 730 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0120] The electronic device 700 can also communicate with one or more external devices 700′ (e.g., a keyboard, a pointing device, a Bluetooth device, etc.), one or more devices that enable a user to interact with the electronic device 700, and / or any device that enables the electronic device 700 to communicate with one or more other computing devices (e.g., a router, a modem, etc.). This communication can occur via an input / output (I / O) interface 750. Furthermore, the electronic device 700 can communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network such as the Internet) via a network adapter 760. The network adapter 760 can communicate with other modules of the electronic device 700 via the bus 730. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with the electronic device 700, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0121] Through the above description of the embodiments, it is easy for those skilled in the art to understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Figure 8 As shown, the technical solution according to the embodiment of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes a number of instructions to enable a computing device (which can be a personal computer, a server, or a network device, etc.) to execute the above method according to the embodiment of the present disclosure.

[0122] The software product may utilize any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection having one or more wires, a portable 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 thereof.

[0123] The computer-readable storage medium may include a data signal propagated in baseband or as part of a carrier wave, wherein the readable program code is carried. The data signal propagated may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or component. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination thereof.

[0124] Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0125] The above-mentioned computer-readable medium carries one or more programs. When the above-mentioned one or more programs are executed by a device, the computer-readable medium implements the following functions: obtaining financial data of multiple historical users, wherein the financial data includes default status; screening multiple positive sample users from the multiple historical users based on the default status; generating a hypersphere equation group through the financial data of the multiple positive sample users and the initial hypersphere equation; solving the hypersphere equation to obtain its optimal solution; and generating a user financial risk identification model based on the optimal solution, wherein the user financial risk identification model is used to determine the user's financial default risk probability.

[0126] Those skilled in the art will appreciate that the modules described above can be distributed in the device according to the description of the embodiment, or can be modified accordingly to be used in one or more devices that are different from the embodiment. The modules of the above embodiment can be combined into one module or further divided into multiple submodules.

[0127] Through the description of the above embodiments, it will be readily understood by those skilled in the art that the example embodiments described herein can be implemented via software or via a combination of software and necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, or mobile hard drive) or on a network and includes several instructions for enabling a computing device (such as a personal computer, server, mobile terminal, or network device) to execute the methods according to the embodiments of the present disclosure.

[0128] While the exemplary embodiments of the present disclosure have been specifically illustrated and described above, it should be understood that the present disclosure is not limited to the detailed structures, configurations, or implementations described herein; rather, the present disclosure is intended to encompass various modifications and equivalent configurations within the spirit and scope of the appended claims.

Claims

1. A method for generating a user financial risk identification model, characterized in that: include: Obtaining financial data of multiple historical users, wherein the financial data includes default status; Filtering a plurality of positive sample users from the plurality of historical users according to the default status, wherein the positive sample set includes financial data of users with a history of default; Classifying the financial data of the plurality of positive sample users according to preset dimensions to generate a plurality of multi-dimensional financial data; Comparing the plurality of multi-dimensional financial data with a preset multi-dimensional feature vector benchmark value to generate multi-dimensional financial data feature values ​​for the plurality of positive sample users, including: comparing the average income of users in the positive sample and determining the feature value of the financial data in that dimension through variance or standard deviation; determining the average income of users in the positive sample set based on the financial data of users in the positive sample set, and then determining the benchmark value; comparing the age of users in the positive sample set with a standard age range to determine the feature value of the age dimension; Generate a hypersphere equation group based on the multi-dimensional financial data feature values ​​of the plurality of positive sample users and the initial hypersphere equation; Determine the target slack variable threshold and optimization objective; Solving the initial hypersphere equation by an optimization algorithm based on the target slack variable threshold and the optimization target to obtain an optimal solution to the initial hypersphere equation; The optimization goal is that all positive sample users are within the hypersphere; generating a target hypersphere equation based on the optimal solution; generating a user financial risk identification model based on the target hypersphere equation, wherein the user financial risk identification model is used to determine the user's financial default risk probability; Inputting the current user's financial data into the target hypersphere equation of the financial risk identification model; The target hypersphere equation is solved using the Lagrangian dual method to obtain the distance between the user's multi-dimensional feature f(x, y, z) and the center of the target hypersphere equation φ(x, y, z) = 0. The Lagrangian function L = f(x, y, z) + λφ(x, y, z) is constructed, where λ is called the Lagrangian multiplier. The partial derivatives of L with respect to x, y, z, and λ are obtained to obtain a system of equations, and the stationary point P(x, y, z) is obtained. The minimum value of the above formula is the center distance. When the sphere center distance is greater than the radius of the target hypersphere, the user is determined to be a low-risk default user; When the sphere center distance is less than or equal to the radius of the target hypersphere, the user is determined to be a high-risk default user.

2. The method according to claim 1, wherein Also includes: Determining characteristic dimensions of the financial data of the historical user; as well as The multi-dimensional feature vector reference value is determined based on the feature dimension and the numerical values ​​of the financial data of multiple historical users corresponding to the dimension.

3. The method according to claim 1, wherein Generating the hypersphere equation group based on the multi-dimensional financial data feature values ​​of the plurality of positive sample users and the initial hypersphere equation includes: Determining a plurality of position coordinates corresponding to the plurality of positive sample users in a multi-dimensional coordinate system based on the multi-dimensional financial data feature values ​​thereof; and The plurality of position coordinates are substituted into the initial hypersphere equations respectively to generate the hypersphere equation group.

4. A user financial risk identification model generation device, characterized in that: include: A data module, configured to obtain financial data of multiple historical users, wherein the financial data includes default status; a screening module, configured to screen out a plurality of positive sample users from the plurality of historical users according to the default status, wherein the positive sample set includes financial data of users with a history of default; an equation module, configured to classify the financial data of the plurality of positive sample users according to preset dimensions to generate a plurality of multi-dimensional financial data; Comparing the plurality of multi-dimensional financial data with a preset multi-dimensional feature vector benchmark value to generate multi-dimensional financial data feature values ​​for the plurality of positive sample users, including: comparing the average income values ​​of users in the positive sample set and determining the feature value of the financial data in that dimension through variance or standard deviation; determining the average income value of users in the positive sample set based on the financial data of users in the positive sample set, and then determining the benchmark value; comparing the age of users in the positive sample set with a standard age range to determine the feature value of the age dimension; and generating a hypersphere equation system based on the multi-dimensional financial data feature values ​​of the plurality of positive sample users and an initial hypersphere equation; A solution module is configured to determine a target slack variable threshold and an optimization target; solve the initial hypersphere equation using an optimization algorithm based on the target slack variable threshold and the optimization target to obtain an optimal solution to the initial hypersphere equation; wherein the optimization target is that all positive sample users are within the hypersphere; A model module, configured to generate a target hypersphere equation based on the optimal solution; and to generate a user financial risk identification model based on the target hypersphere equation, wherein the user financial risk identification model is used to determine the user's financial default risk probability; An input module, configured to input the current user's financial data into the target hypersphere equation of the financial risk identification model; The distance module is used to calculate the distance between the multi-dimensional feature f(x, y, z) and the center of the target hypersphere equation φ(x, y, z) = 0. The Lagrangian function L = f(x, y, z) + λφ(x, y, z) is used, where λ is called the Lagrangian multiplier. The partial derivatives of L with respect to x, y, z, and λ are obtained to obtain the system of equations and the stationary point P(x, y, z). The minimum value of the above formula is the center distance. The risk module is configured to determine that the user is a low-risk default user when the sphere center distance is greater than the target hypersphere radius; and to determine that the user is a high-risk default user when the sphere center distance is less than or equal to the target hypersphere radius.

5. The device according to claim 4, characterized in that The equation module also includes a feature unit, which includes: a dimension subunit, configured to determine characteristic dimensions of the financial data of the historical user; and The benchmark subunit is configured to determine a benchmark value of the multi-dimensional feature vector based on the feature dimension and the numerical value of the financial data of multiple historical users corresponding to the dimension.

6. The device according to claim 4, characterized in that The equation module further includes an equation group unit, wherein the equation group unit includes: a coordinate subunit, configured to determine a plurality of position coordinates corresponding to the plurality of positive sample users in a multi-dimensional coordinate system based on the multi-dimensional financial data feature values ​​of the plurality of positive sample users; and The substitution subunit is used to substitute the multiple position coordinates into the initial hypersphere equations respectively to generate the hypersphere equation group.

7. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 3.

8. A computer-readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 3 is implemented.

Images