Credit business data processing method and device based on artificial intelligence

Abstract

The present invention provides a method and device for processing credit business data based on artificial intelligence, which relates to the field of intelligent decision-making technology and the financial field. The method includes: extracting risk factors corresponding to the credit business to be approved; determining the risk level corresponding to the credit business to be approved based on the risk factors and a preset risk mapping relationship; and approving the credit business to be approved according to the approval method corresponding to the risk level to obtain an approval processing result. The artificial intelligence-based credit business data processing method and device of the present application improves computing efficiency by distinguishing business risk types, adopting a more accurate calculation method for high-risk businesses, and adopting a computing model with a simple algorithm and small amount of computing for low-risk businesses, thereby reducing the amount of system computing and improving the operating efficiency of the automatic approval system.

Description

Technical Field

[0001] The present application relates to the fields of intelligent decision-making technology and finance, and specifically to a credit business data processing method and device based on artificial intelligence. Background Art

[0002] Nowadays, automated approval technology is being used more and more widely in the field of bank credit business. The application of automated approval technology has effectively reduced the manual workload of credit approval personnel.

[0003] However, as credit business volume gradually increases, existing automated approval methods mostly build relatively complex system algorithm models, performing indiscriminate approval operations regardless of the business data. This increases the amount of data processed during the approval process, and also reduces the efficiency and success rate of the automated approval system. Therefore, how to achieve automated approval while improving approval efficiency is an urgent problem to be solved. Summary of the Invention

[0004] In order to improve the efficiency of automated approval of credit business, in the first aspect, this application provides a credit business data processing method based on artificial intelligence, which relates to the field of intelligent decision-making technology and the financial field. The method includes:

[0005] Extract risk factors corresponding to credit business awaiting approval;

[0006] Determine the risk level corresponding to the credit business to be approved based on the risk factors and the preset risk mapping relationship;

[0007] The credit business to be approved is approved according to the approval method corresponding to the risk level to obtain an approval result.

[0008] In one embodiment, determining the risk level corresponding to the credit business to be approved based on the risk factors and a preset risk mapping relationship includes:

[0009] Obtain the corresponding element values ​​of the customer rating, guarantee type, business type and margin in the risk factors;

[0010] The risk level of the credit business to be approved is determined according to the factor values ​​and the risk mapping relationship, wherein the risk mapping relationship includes the risk level corresponding to the value combination of each risk factor.

[0011] In one embodiment, when the risk level is a high risk level, the approval process for the credit business to be approved is performed according to the approval method corresponding to the risk level to obtain an approval process result, including:

[0012] Capture the approval elements of the business to be approved;

[0013] Determining an approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and a preset approval factor assignment relationship;

[0014] The approval processing result is obtained according to the approval vector and a vector matrix of the customer corresponding to the credit business to be approved, wherein the vector matrix is ​​obtained according to the historical business data of the customer corresponding to the credit business to be approved.

[0015] In one embodiment, obtaining the approval processing result according to the approval vector and the vector matrix of the customer corresponding to the credit business to be approved includes:

[0016] Performing Schmidt orthogonalization and normalization on the vector matrix to obtain a standard orthogonal vector group;

[0017] Performing an inner product operation on the approval vector and the standard orthogonalized vector group to obtain an inner product result;

[0018] The approval processing result is determined according to the inner product result.

[0019] In one embodiment, the step of obtaining the vector matrix includes:

[0020] Obtain multiple historical business data of customers corresponding to pending credit business;

[0021] Capture the approval elements of each piece of historical business data separately;

[0022] Assigning values ​​to the approval factors according to the approval factors and the preset approval factor assignment relationship to obtain an approval vector corresponding to each piece of historical business data;

[0023] The vector matrix is ​​generated according to the approval vector corresponding to each historical business data.

[0024] In one embodiment, when the risk level is a low risk level, the approval process for the credit business to be approved is performed according to the approval method corresponding to the risk level, and the approval process result is obtained, including:

[0025] Capture the approval elements of the business to be approved;

[0026] Determining an approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and a preset approval factor assignment relationship;

[0027] The approval processing result is obtained according to the approval vector and the historical vector of the recent historical business data of the customer corresponding to the credit business to be approved. The historical vector is determined according to the approval factors of the historical business data and the preset approval factor assignment relationship.

[0028] In one embodiment, obtaining the approval processing result based on the approval vector and the historical vector of the most recent historical business data of the customer corresponding to the credit business to be approved includes:

[0029] Performing an angle operation on the approval vector and the history vector to obtain a cosine result of the angle;

[0030] The approval processing result is determined according to the cosine result of the angle.

[0031] In a second aspect, the present application provides an artificial intelligence-based credit business data processing device, comprising:

[0032] Factor extraction module, used to extract risk factors corresponding to pending credit business;

[0033] A risk level classification module is used to determine the risk level corresponding to the credit business to be approved based on the risk factors and the preset risk mapping relationship;

[0034] The approval module is used to approve the credit business according to the approval method corresponding to the risk level and obtain the approval result.

[0035] In one embodiment, the risk level classification module includes:

[0036] An element value determination unit is used to obtain the element values ​​corresponding to the customer rating, guarantee type, business type and margin in the risk factors;

[0037] The risk level determination unit is used to determine the risk level of the credit business to be approved based on the factor values ​​and the risk mapping relationship, where the risk mapping relationship includes the risk levels corresponding to the value combinations of the risk factors.

[0038] In one embodiment, when the risk level is a high risk level, the approval module includes:

[0039] Approval element capture unit, used to capture the approval elements of the business to be approved;

[0040] An approval vector determining unit, configured to determine an approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and a preset approval factor assignment relationship;

[0041] The approval processing unit is used to obtain the approval processing result according to the approval vector and the vector matrix of the customer corresponding to the credit business to be approved, wherein the vector matrix is ​​obtained according to the historical business data of the customer corresponding to the credit business to be approved.

[0042] In one embodiment, the approval processing unit is specifically configured to:

[0043] Performing Schmidt orthogonalization and normalization on the vector matrix to obtain a standard orthogonal vector group;

[0044] Performing an inner product operation on the approval vector and the standard orthogonalized vector group to obtain an inner product result;

[0045] The approval processing result is determined according to the inner product result.

[0046] In one embodiment, the artificial intelligence-based credit business data processing device further includes a vector matrix determination module for:

[0047] Obtain multiple historical business data of customers corresponding to pending credit business;

[0048] Capture the approval elements of each piece of historical business data separately;

[0049] Assigning values ​​to the approval factors according to the approval factors and the preset approval factor assignment relationship to obtain an approval vector corresponding to each piece of historical business data;

[0050] The vector matrix is ​​generated according to the approval vector corresponding to each historical business data.

[0051] In one embodiment, when the risk level is a low risk level, the approval module includes:

[0052] Approval element capture unit, used to capture the approval elements of the business to be approved;

[0053] An approval vector determining unit, configured to determine an approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and a preset approval factor assignment relationship;

[0054] The approval processing unit is used to obtain the approval processing result based on the approval vector and the historical vector of the recent historical business data of the customer corresponding to the credit business to be approved, wherein the historical vector is determined based on the approval factors of the historical business data and the preset approval factor assignment relationship.

[0055] In one embodiment, the approval processing unit is specifically configured to:

[0056] Performing an angle operation on the approval vector and the history vector to obtain a cosine result of the angle;

[0057] The approval processing result is determined according to the cosine result of the angle.

[0058] In a third aspect, the present application provides an electronic device, comprising:

[0059] A central processing unit, a memory, and a communication module, wherein the memory stores a computer program, the central processing unit can call the computer program, and when the central processing unit executes the computer program, it implements any artificial intelligence-based credit business data processing method provided in this application.

[0060] In a fourth aspect, the present application provides a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements any artificial intelligence-based credit business data processing method provided in the present application.

[0061] The artificial intelligence-based credit business data processing method and device of the present application improves computing efficiency by distinguishing business risk types. High-risk businesses adopt a more accurate calculation method, and the amount of calculation at one time is n vector multiplications. Low-risk businesses adopt an angle calculation model with a simple algorithm and small amount of calculation, and the amount of calculation at one time is 1 vector multiplication, thereby reducing the system's amount of calculation and improving the operating efficiency of the automatic approval system. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0063] Figure 1 A schematic diagram of the artificial intelligence-based credit business data processing method provided for this application.

[0064] Figure 2 Schematic diagram of the steps for determining the risk level of a pending credit application provided for this application.

[0065] Figure 3 Schematic diagram of the approval process steps for high-risk pending credit business provided for this application.

[0066] Figure 4 Schematic diagram of the steps for generating a vector matrix provided in this application.

[0067] Figure 5 A schematic diagram of the steps for determining the approval processing results of high-risk pending credit business provided for this application.

[0068] Figure 6 A flowchart for determining the approval processing results based on the inner product results provided for this application.

[0069] Figure 7 Schematic diagram of the approval process steps for low-risk pending credit business provided for this application.

[0070] Figure 8 A schematic diagram of the steps for determining the approval processing results of low-risk pending credit business provided for this application.

[0071] Figure 9 A schematic diagram of the artificial intelligence-based credit business data processing device provided in this application.

[0072] Figure 10 Another schematic diagram of the artificial intelligence-based credit business data processing device provided for this application.

[0073] Figure 11 Another schematic diagram of the artificial intelligence-based credit business data processing device provided for this application.

[0074] Figure 12 Another schematic diagram of the artificial intelligence-based credit business data processing device provided for this application.

[0075] Figure 13 Another schematic diagram of the artificial intelligence-based credit business data processing device provided for this application.

[0076] Figure 14 A schematic diagram of an electronic device provided in this application. DETAILED DESCRIPTION

[0077] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0078] In order to improve the efficiency of automated approval of credit business, on the first hand, this application provides a credit business data processing method based on artificial intelligence, and the execution entity of this method can be a banking institution or other financial institution that provides credit business to the outside world. First of all, it should be noted that the acquisition, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations. The user information in the embodiments of this application is obtained through legal and compliant channels, and the acquisition, storage, use, and processing of user information are authorized and agreed by the customer.

[0079] like Figure 1 As shown, the method includes steps S101 to S103:

[0080] Step S101: extract risk factors corresponding to the credit business to be approved.

[0081] Specifically, the risk factors in this step refer to the data elements within the business data of the pending credit business that can be used to determine the potential risk of the customer applying for the credit business. These include, but are not limited to, customer rating, guarantee type, business type, margin, and other data items. A comprehensive assessment of the customer's potential risk based on these data items can yield a more objective assessment of the customer's risk level.

[0082] In actual applications, risk factors can be flexibly adjusted according to the purpose of data processing or the bank's focus, and it is not necessary to use the data items listed in this step.

[0083] Step S102: determining the risk level corresponding to the credit business to be approved based on the risk factors and a preset risk mapping relationship.

[0084] This step classifies the risk levels of pending credit transactions into two categories: high-risk and low-risk. The risk mapping relationship specifies the risk level classification results corresponding to different combinations of risk factor values ​​in the pending credit transactions. This step gives a possible business risk level classification rule: if the customer rating is high, and the guarantee type is high-quality guarantee or the business type is low-risk business type or the margin is full margin, then the credit business to be approved is low-risk business; if the customer rating is high, and at the same time meets the three conditions of the guarantee type is not high-quality guarantee, the business type is not low-risk business type, and the margin is not full margin, then the credit business to be approved is high-risk business; if the customer rating is low, and meets at least two of the three conditions of the guarantee type is high-quality guarantee, the business type is low-risk business type, and the margin is full margin, then the credit business to be approved is low-risk business; if the customer rating is low, and only meets one of the three conditions of the guarantee type is high-quality guarantee, the business type is low-risk business type, and the margin is full margin, or none of the three conditions of the guarantee type is high-quality guarantee, the business type is low-risk business type, and the margin is full margin are met, then the credit business to be approved is high-risk business.

[0085] The risk mapping relationship may be stored in a database or other data storage unit / device in the form of a risk mapping relationship table.

[0086] Step S103: Approving the credit business according to the approval method corresponding to the risk level to obtain an approval result.

[0087] This application uses different approval methods for automated approval of credit transactions at different risk levels. This step retrieves the corresponding approval method for the credit transaction based on its risk level. Approval is performed according to the obtained approval method to obtain the corresponding approval result. There are two types of approval results: approved or rejected.

[0088] This embodiment improves computing efficiency by differentiating business risk types. High-risk businesses use more accurate calculation methods, while low-risk businesses use calculation models with simple algorithms and small computational load. This reduces the system's computational load and improves the operating efficiency of the automatic approval system.

[0089] In one embodiment, if Figure 2 As shown, step S102, determining the risk level corresponding to the credit business to be approved based on the risk factors and the preset risk mapping relationship, includes:

[0090] Step S1021, obtaining the corresponding element values ​​of the customer rating, guarantee type, business type and margin in the risk factors.

[0091] Specifically, customer ratings may include, for example, AAA+, AAA, AA+, AA, A+, A, B+, B, C+, C, etc., where a rating of AAA+, AAA, AA+, or AA indicates that the customer is a high-rated customer, and otherwise, the customer is a low-rated customer. The customer rating of this application is a result data obtained directly from the corresponding end of the executing entity. The steps for determining the customer rating are not performed by this application, and therefore this application does not limit the process of determining the customer rating.

[0092] For example, the guarantee type may include high-quality guarantee and non-high-quality guarantee; the business types may include low-risk business types, medium-risk business types, and high-risk business types; and the margin may include full margin and part-full margin.

[0093] Step S1022: determining the risk level of the credit business to be approved based on the factor values ​​and the risk mapping relationship, wherein the risk mapping relationship includes the risk levels corresponding to the value combinations of the risk factors.

[0094] Based on the specific element values ​​of each risk factor given in this embodiment and the risk mapping relationship given in step S102 of the previous embodiment, several specific examples are provided here to show the risk element values ​​of the pending credit business and the corresponding risk level determination results, as shown in Table 1 below:

[0095] Table 1: Risk level determination results

[0096]

[0097] In one embodiment, if Figure 3 As shown, when the risk level is a high risk level, step S103 is to process the credit business to be approved according to the approval method corresponding to the risk level, and obtain the approval processing result, including:

[0098] Step S1031: Capture the approval elements of the business to be approved.

[0099] Specifically, approval factors include but are not limited to loan amount, loan term, customer rating, loan type, and repayment method.

[0100] Step S1032: determining an approval vector corresponding to the credit business to be approved according to the approval factors of the credit business to be approved and a preset approval factor assignment relationship.

[0101] Specifically, the evaluation rules corresponding to each approval factor are specified in the evaluation relationship of the approval factors. For example, for the loan amount, if the loan amount is within the range of 0 to 10 million, the loan amount factor is assigned a value of 1; if the loan amount is within the range of 10 million to 100 million, the loan amount factor is assigned a value of 2; if the loan amount is above 100 million, the loan amount factor is assigned a value of 3; for the loan term, if the loan term is within 6 months, the loan term factor is assigned a value of 1; if the loan term is between 6 months and 1 year, the loan term factor is assigned a value of 2; if the loan term is above 1 year, the loan term factor is assigned a value of 3; for the customer rating (see step S1 in the above embodiment), 021), if the customer rating is AA or above, the customer rating factor is assigned a value of 1; if the customer rating is A+ or below, the customer rating factor is assigned a value of 2; for loan types, if the loan type is a proprietary loan, the loan type factor is assigned a value of 1; if the loan type is an entrusted loan, the loan type factor is assigned a value of 2; if the loan type is a specific loan, the loan type factor is assigned a value of 3; for repayment method, if the repayment method is equal installments of principal and interest or equal installments of principal, the repayment method factor is assigned a value of 1; if the repayment method is other repayment methods, such as equal incremental (decreasing) or scheduled interest and principal repayment, the repayment method factor is assigned a value of 2.

[0102] The approval factor assignment relationship can be stored in a database or other data storage unit / device in the form of an approval factor assignment relationship table, as shown in Table 2:

[0103] Table 2: Approval element assignment relationship table

[0104]

[0105]

[0106] Those skilled in the art will appreciate that the approval factors and their value assignments provided in this embodiment are examples provided by this application for illustrative purposes only and are not intended to limit this application. In actual applications, more or fewer pieces of credit business data to be approved may be selected as approval factors, and the value assignments may be flexibly adjusted based on relevant regulations and the credit market situation. This application does not impose any limitations on these.

[0107] After determining the values ​​of each approval factor according to the approval factor assignment relationship, the approval vector is formed based on each assignment. For example, suppose the values ​​of n approval factors such as loan amount, loan term, customer rating, loan type, repayment method, etc. of the credit business to be approved are 1, 3, 2, 1, 2, ..., x respectively. n , then the approval vector corresponding to the credit business to be approved is expressed as α=[a1,a2,a3,a4,a5,…,a n ]=[1,3,2,1,2,…,x n ], where α i Represents the i-th element in the approval vector α, 1≤i≤n, and n is a positive integer greater than 1.

[0108] Step S1033 , obtaining the approval processing result according to the approval vector and the vector matrix of the customer corresponding to the credit business to be approved, wherein the vector matrix is ​​obtained according to the historical business data of the customer corresponding to the credit business to be approved.

[0109] The approval vector is the vector α obtained in step S1032 = [1, 3, 2, 1, 2, ..., x n ], which is transposed into a column vector:

[0110]

[0111] The vector matrix of customers corresponding to the pending credit business with a high risk level is generated based on multiple pieces of historical business data of the customers corresponding to the pending credit business.

[0112] Specifically, the steps of generating the vector matrix can be found in Figure 4 , including the following steps:

[0113] Step S401: Acquire multiple pieces of historical business data of customers corresponding to the credit business to be approved.

[0114] The multiple pieces of historical business data in this step may be the historical business data of all loans already disbursed to the customer corresponding to the credit business to be approved.

[0115] Step S402: Capture the approval elements of each piece of historical business data.

[0116] Specifically, for each piece of historical business data, we extract approval factors such as loan amount, loan term, customer rating, loan type, and repayment method;

[0117] Step S403: Assign values ​​to the approval elements according to the approval elements and the preset approval element assignment relationship to obtain an approval vector corresponding to each piece of historical business data.

[0118] Specifically, the relationship between the approval elements and the method of assigning values ​​to each approval element can be found in the description of step S1032 of the above embodiment, and will not be repeated here. The approval vectors corresponding to each historical business data can be expressed as:

[0119] p1=[p 11 , p 12 , p 13 ,…,p 1i ,…,p 1n ], where p 1i represents the i-th element in the approval vector p1, 1≤i≤n, and n is a positive integer greater than 1;

[0120] p2=[p 21 , p 22 , p 23 ,…,p 2i ,…,p 2n ], where p 2i represents the i-th element in the approval vector p2, 1≤i≤n, and n is a positive integer greater than 1;

[0121] p3=[p 31 , p 32 , p 33 ,…,p 3i ,…,p 3n ], where p 3i represents the i-th element in the approval vector p3, 1≤i≤n, and n is a positive integer greater than 1;

[0122] …

[0123] Until the approval vectors p1-pm corresponding to each piece of historical business data are obtained, where m is a positive integer greater than 1, representing the total number of historical business data.

[0124] Step S404: Generate the vector matrix according to the approval vector corresponding to each historical business data.

[0125] Specifically, the approval vectors corresponding to the historical business data are combined into an m×n vector matrix P:

[0126]

[0127] Among them, P is a vector matrix, p1~pm are the approval vectors corresponding to each historical business data, p ji is the i-th element in the j-th approval vector pj, 1≤j≤m, m is a positive integer greater than 1, 1≤i≤n, n is a positive integer greater than 1.

[0128] In one embodiment, if Figure 5 As shown, step S1033, obtaining the approval processing result according to the approval vector and the vector matrix of the customer corresponding to the credit business to be approved, includes:

[0129] Step S10331, performing Schmidt orthogonalization and normalization on the vector matrix to obtain a standard orthogonal vector group.

[0130] Specifically, each row vector in the vector matrix P is Schmidt orthogonalized and normalized to obtain a standard orthogonal vector set β = {β1, β2, β3, ..., β j ,…,β m}.

[0131] Step S10332: Perform an inner product operation on the approval vector and the standard orthogonalized vector group to obtain an inner product result.

[0132] Specifically, the inner product operation is performed on the approval vector α and each vector in the standard orthogonalized vector group β:

[0133] A j =β j α T =β j1 a1+β j2 a2+β j3 a3+……+β ji a i +……+β jn a n

[0134] Among them, A j is the vector β j With vector α T The inner product operation result, β ji is the vector β j The i-th element in a i is the vector α T The i-th element in , 1≤j≤m, m is a positive integer greater than 1, 1≤i≤n, n is a positive integer greater than 1.

[0135] This step obtains m inner product results, namely A1, A2, ..., A j ,……,A m .

[0136] Step S10333: Determine the approval processing result according to the inner product result.

[0137] Specifically, the number of 0s in the m inner product results is counted. If all m inner product results are 0, the approval processing result is determined to be disapproval; if the number of 0s in the inner product results is greater than or equal to half of the total number of inner product results and less than the total number of inner product results m, the approval processing result is determined to be returning to the business investigation post for supplementary information; if the number of 0s in the inner product results is less than half of the total number of inner product results, the approval processing result is determined to be approval.

[0138] The above steps can be performed, for example, Figure 6 The flowchart shown is used for statistics:

[0139] (1) Assign initial values ​​j = 1, k = 0; where j is the inner product result A j The subscript of , 1≤j≤m, m is a positive integer greater than 1, and k is the number of 0s in the inner product result;

[0140] (2) Judge A j Is it equal to 0? If so, go to step (3), if not, go to step (4);

[0141] (3) Let k = k + 1, that is, the number of 0s plus 1;

[0142] (4) Let j = j + 1, that is, to determine the next inner product result;

[0143] (5) Determine whether j>m holds; if so, all inner product results have been determined, and step (6) is executed; if not, return to step (2);

[0144] (6) Determine whether k=m; if so, execute step (8); if not, execute step (7);

[0145] (7)Judgment Is it true? If so, go to step (9); if not, go to step (10);

[0146] (8) Output “Approval failed” and end.

[0147] (9) Output “Return to Business Investigation Post for Supplementary Information” and end.

[0148] (10) Output “Approved” and end.

[0149] In one embodiment, if Figure 7 As shown, when the risk level is low, step S103 is to process the credit business to be approved according to the approval method corresponding to the risk level, and obtain the approval processing result, including:

[0150] Step S1034: Capture the approval elements of the business to be approved.

[0151] Step S1034 is similar to step S1031 , so for a more specific description of step S1034 in this embodiment, please refer to the description of step S1031 , which will not be repeated here.

[0152] Step S1035 : determining an approval vector corresponding to the credit business to be approved according to the approval factors of the credit business to be approved and a preset approval factor assignment relationship.

[0153] Step S1035 is similar to step S1032, so for a more specific description of step S1035 in this embodiment, please refer to the description of step S1032, which will not be repeated here.

[0154] Step S1036, obtaining the approval processing result according to the approval vector and the historical vector of the recent historical business data of the customer corresponding to the credit business to be approved, wherein the historical vector is determined according to the approval factors of the historical business data and the preset approval factor assignment relationship.

[0155] Assume that the approval vector obtained in this step is γ = [b1, b2, b3, b4, b5, ..., b n ]=[3,1,1,2,2,…,y n ], where b i Represents the i-th element in the approval vector γ, 1≤i≤n, and n is a positive integer greater than 1. The transpose of the vector γ is a column vector:

[0156]

[0157] The historical vector is determined based on the approval factors of the historical business data and the preset approval factor assignment relationship. The approval factors and the approval factor assignment relationship are described in steps S1031 and S1032 and will not be repeated here. After obtaining the most recent historical business data of the credit business to be approved, the approval factors of this historical business data are captured and assigned according to steps S1031 and S1032 (or steps S1034 and S1035), and the vector model Q corresponding to this historical business data is obtained:

[0158] Q=(q1, q2, q3,..., q f ,……,q n )

[0159] Among them, q f Represents the f-th element in the approval vector Q, 1≤f≤n, and n is a positive integer greater than 1.

[0160] As can be seen from the above, for pending credits with a low risk level, it is not necessary to obtain multiple historical business data (all historical business data for disbursed loans) as is done in step S1033 for pending credits with a high risk level. Instead, only the most recent historical business data is required. This difference reflects the significant difference in the amount of data computation required during the automated approval process for pending credits with a low risk level and a high risk level. This helps improve the efficiency of automated approval while ensuring the reliability of risk approval.

[0161] In one embodiment, if Figure 8 As shown, step S1036, obtaining the approval processing result according to the approval vector and the recent historical business data of the customer corresponding to the credit business to be approved, includes:

[0162] Step S10361: perform angle calculation on the approval vector and the history vector to obtain a cosine result of the angle.

[0163] Specifically, the cosine of the angle between the approval vector γ and the history vector Q is calculated according to the following formula:

[0164]

[0165] in,

[0166] Q·γ=q1b1+q2b2+……+q f b i +……+q n b n

[0167]

[0168]

[0169] q f represents the fth element in the approval vector Q, 1≤f≤n; b i Represents the i-th element in the approval vector γ, n is a positive integer greater than 1, 1≤i≤n, and n is a positive integer greater than 1.

[0170] The angle between the approval vector γ and the history vector Q can be further obtained by the following formula:

[0171] θ=arc cosθ

[0172] Step S10362: Determine the approval processing result based on the cosine result of the angle.

[0173] Specifically, if the angle cosine result cosθ is in the range of [0,1], it means that the approval vector γ and the historical vector Q have a high similarity, and the approval processing result is determined to be approved; when the angle cosine is in the range of [-1, 0), it means that the approval vector γ and the historical vector Q have a low similarity, and the approval processing result is determined to be unapproved.

[0174] For low-risk pending credit transactions, this method only requires a single cosine operation of the angle between the approval vector γ and the history vector Q to arrive at the approval result. For high-risk pending credit transactions, the approval result requires multiplying each of the n vectors in the standard orthogonal vector set β corresponding to the vector matrix P by the approval vector, resulting in n operations. Therefore, the computational complexity of this method for approving low-risk pending credit transactions is much lower than that of the high-risk pending credit transactions.

[0175] It can be seen that the artificial intelligence-based credit business data processing method of this application improves computing efficiency by distinguishing business risk types. High-risk businesses adopt a more accurate calculation method, and the amount of calculation at one time is n vector multiplications. Low-risk businesses adopt an angle calculation model with a simple algorithm and small amount of calculation, and the amount of calculation at one time is 1 vector multiplication, thereby reducing the system's amount of calculation, saving calculation time, and improving the operating efficiency of the automatic approval system.

[0176] Based on the same inventive concept, the embodiments of the present application also provide an artificial intelligence-based credit business approval device, which can be used to implement the method described in the above embodiments, as described in the following embodiments. Since the principle of solving the problem by the artificial intelligence-based credit business approval device is similar to that of the artificial intelligence-based credit business approval method, the implementation of the artificial intelligence-based credit business approval device can refer to the implementation of the artificial intelligence-based credit business approval method, and the repetitions will not be repeated. As used below, the term "unit" or "module" can be a combination of software and / or hardware that implements a predetermined function. Although the system described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and conceived.

[0177] like Figure 9 As shown, the present application provides a credit business data processing device based on artificial intelligence, comprising:

[0178] Factor extraction module 901, used to extract risk factors corresponding to the credit business to be approved;

[0179] The risk level classification module 902 is used to determine the risk level corresponding to the credit business to be approved based on the risk factors and the preset risk mapping relationship;

[0180] The approval module 903 is used to approve the credit business according to the approval method corresponding to the risk level and obtain the approval result.

[0181] In one embodiment, if Figure 10 As shown, the risk level classification module 902 includes:

[0182] The element value determination unit 9021 is used to obtain the element values ​​corresponding to the customer rating, guarantee type, business type and margin in the risk factors;

[0183] The risk level determination unit 9022 is configured to determine the risk level of the credit business to be approved based on the factor values ​​and the risk mapping relationship, wherein the risk mapping relationship includes the risk levels corresponding to the value combinations of the risk factors.

[0184] In one embodiment, if Figure 11 As shown, the approval module 903 includes:

[0185] The first approval factor capturing unit 9031 is used to capture the approval factors of the business to be approved when the risk level is a high risk level;

[0186] A first approval vector determining unit 9032 is configured to determine an approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and a preset approval factor assignment relationship when the risk level is a high risk level;

[0187] The first approval processing unit 9033 is used to obtain the approval processing result based on the approval vector and the vector matrix of the customer corresponding to the credit business to be approved when the risk level is a high risk level, wherein the vector matrix is ​​obtained based on the historical business data of the customer corresponding to the credit business to be approved.

[0188] In one embodiment, the first approval processing unit 9033 is specifically configured to:

[0189] Performing Schmidt orthogonalization and normalization on the vector matrix to obtain a standard orthogonal vector group;

[0190] Performing an inner product operation on the approval vector and the standard orthogonalized vector group to obtain an inner product result;

[0191] The approval processing result is determined according to the inner product result.

[0192] In one embodiment, if Figure 12 As shown, the artificial intelligence-based credit business data processing device further includes a vector matrix determination module 904, which is used to:

[0193] Obtain multiple historical business data of customers corresponding to pending credit business;

[0194] Capture the approval elements of each piece of historical business data separately;

[0195] Assigning values ​​to the approval factors according to the approval factors and the preset approval factor assignment relationship to obtain an approval vector corresponding to each piece of historical business data;

[0196] The vector matrix is ​​generated according to the approval vector corresponding to each historical business data.

[0197] In one embodiment, if Figure 13 As shown, the approval module 903 also includes:

[0198] The second approval factor capturing unit 9034 is used to capture the approval factors of the business to be approved when the risk level is low;

[0199] The second approval vector determining unit 9035 is configured to determine the approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and the preset approval factor assignment relationship when the risk level is a low risk level;

[0200] The second approval processing unit 9036 is used to obtain the approval processing result based on the approval vector and the historical vector of the recent historical business data of the customer corresponding to the credit business to be approved when the risk level is a low risk level. The historical vector is determined based on the approval factors of the historical business data and the preset approval factor assignment relationship.

[0201] In one embodiment, the second approval processing unit 9036 is specifically configured to:

[0202] Performing an angle operation on the approval vector and the history vector to obtain a cosine result of the angle;

[0203] The approval processing result is determined according to the cosine result of the angle.

[0204] The artificial intelligence-based credit business data processing device of the present application improves computing efficiency by distinguishing business risk types. A more accurate calculation method is adopted for high-risk businesses, and the amount of calculation at one time is n vector multiplications. A simple algorithm and a small amount of calculation are adopted for low-risk businesses. The amount of calculation at one time is only one vector multiplication, thereby reducing the amount of system calculation and improving the operating efficiency of the automatic approval system.

[0205] In a third aspect, the present invention further provides an electronic device, see Figure 14 , the electronic device 100 specifically includes:

[0206] A central processing unit (CPU) 110 , a memory (memory) 120 , a communication module (Communications) 130 , an input unit 140 , an output unit 150 and a power supply 160 .

[0207] The memory 120, communication module 130, input unit 140, output unit 150, and power supply 160 are respectively connected to the central processing unit 110. The memory 120 stores a computer program, which can be called by the central processing unit 110. When the central processing unit 110 executes the computer program, all steps of the artificial intelligence-based credit business data processing method in the above embodiment are implemented.

[0208] In a fourth aspect, embodiments of the present application further provide a computer-readable storage medium for storing a computer program, wherein the computer program is executable by a processor. When the computer program is executed by the processor, the computer program implements any of the artificial intelligence-based credit business data processing methods provided by the present invention.

[0209] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0210] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0211] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0212] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0213] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the embodiments of this specification.

[0214] In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples, unless they contradict each other. The above is only an embodiment of the embodiment of this specification and is not intended to limit the embodiment of this specification. For those skilled in the art, the embodiment of this specification may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiment of this specification shall be included within the scope of the claims of the embodiment of this specification.

Claims

1. A credit business data processing method based on artificial intelligence, characterized in that: include: Extract risk factors corresponding to credit business awaiting approval; Determine the risk level corresponding to the credit business to be approved based on the risk factors and the preset risk mapping relationship; Approving the credit business according to the approval method corresponding to the risk level to obtain the approval result; Wherein, when the risk level is a high risk level, the approval process for the credit business to be approved is performed according to the approval method corresponding to the risk level, and the approval process result is obtained, including: Capture the approval elements of the business to be approved; Determining an approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and a preset approval factor assignment relationship; Perform Schmidt orthogonalization and unitization on the vector matrix to obtain a standard orthogonalized vector group; Performing an inner product operation on the approval vector and the standard orthogonalized vector group to obtain an inner product result; Determining the approval processing result based on the inner product result, wherein the vector matrix is ​​obtained based on multiple historical business data of the customer corresponding to the credit business to be approved; Wherein, when the risk level is a low risk level, the credit business to be approved is approved according to the approval method corresponding to the risk level, and the approval processing result is obtained, including: Capture the approval elements of the business to be approved; Determining an approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and a preset approval factor assignment relationship; Performing an angle calculation on the approval vector and the history vector to obtain a cosine result of the angle; The approval processing result is determined according to the cosine result of the angle, wherein the history vector is determined according to the approval factors of the historical business data and a preset approval factor assignment relationship.

2. The artificial intelligence-based credit business data processing method according to claim 1, characterized in that: Determining the risk level corresponding to the pending credit business based on the risk factors and the preset risk mapping relationship includes: Obtain the corresponding element values ​​of the customer rating, guarantee type, business type and margin in the risk factors; The risk level of the credit business to be approved is determined according to the factor values ​​and the risk mapping relationship, wherein the risk mapping relationship includes the risk level corresponding to the value combination of each risk factor.

3. The artificial intelligence-based credit business data processing method according to claim 1, characterized in that: The step of obtaining the vector matrix includes: Obtain multiple historical business data of customers corresponding to pending credit business; Capture the approval elements of each piece of historical business data separately; Assigning values ​​to the approval factors according to the approval factors and the preset approval factor assignment relationship to obtain an approval vector corresponding to each piece of historical business data; The vector matrix is ​​generated according to the approval vector corresponding to each historical business data.

4. A credit business data processing device based on artificial intelligence, characterized in that: include: Factor extraction module, used to extract risk factors corresponding to pending credit business; A risk level classification module is used to determine the risk level corresponding to the credit business to be approved based on the risk factors and the preset risk mapping relationship; An approval module is used to approve the credit business according to the approval method corresponding to the risk level and obtain the approval result; The approval module includes: A first approval factor capturing unit, for capturing the approval factors of the business to be approved when the risk level is a high risk level; a first approval vector determining unit, for determining, when the risk level is a high risk level, an approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and a preset approval factor assignment relationship; a first approval processing unit, configured to, when the risk level is a high risk level, perform Schmidt orthogonalization and unitization on a vector matrix to obtain a standard orthogonalized vector group; perform an inner product operation on the approval vector and the standard orthogonalized vector group to obtain an inner product result; and determine the approval processing result based on the inner product result, wherein the vector matrix is ​​obtained based on a plurality of historical business data corresponding to the customer of the credit business to be approved; A second approval factor capturing unit, for capturing the approval factors of the business to be approved when the risk level is a low risk level; a second approval vector determining unit, for determining, when the risk level is a low risk level, an approval vector corresponding to the credit business to be approved based on the approval factors of the credit business to be approved and a preset approval factor assignment relationship; The second approval processing unit is used to perform an angle operation on the approval vector and the historical vector to obtain a cosine result of the angle when the risk level is a low risk level; and determine the approval processing result based on the cosine result of the angle, wherein the historical vector is determined based on the approval factors of the historical business data and the preset approval factor assignment relationship.

5. An electronic device, characterized in that: include: A central processing unit, a memory, and a communication module, wherein the memory stores a computer program, the central processing unit can call the computer program, and when the central processing unit executes the computer program, it implements the credit business data processing method based on artificial intelligence as described in any one of claims 1 to 3.

6. A computer-readable storage medium for storing a computer program, characterized in that: When the computer program is executed by a processor, the credit business data processing method based on artificial intelligence as described in any one of claims 1 to 3 is implemented.

7. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the credit business data processing method based on artificial intelligence as described in any one of claims 1 to 3 are implemented.

Images