CN101996381A - Method and system for calculating retail asset risk - Google Patents
Method and system for calculating retail asset risk Download PDFInfo
- Publication number
- CN101996381A CN101996381A CN2009100909367A CN200910090936A CN101996381A CN 101996381 A CN101996381 A CN 101996381A CN 2009100909367 A CN2009100909367 A CN 2009100909367A CN 200910090936 A CN200910090936 A CN 200910090936A CN 101996381 A CN101996381 A CN 101996381A
- Authority
- CN
- China
- Prior art keywords
- credit
- data
- risk
- retail
- scoring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention provides a method and system for calculating retail asset risk. The method comprises the following steps of: collecting source data, wherein the source data comprises loan and credit card transaction data and customer asset data collected from each business system of a bank, customer and account data collected from a personal credit system and a bank card system in real time during application and customer credit information data collected from a personal credit information system in real time during application; receiving the source data, and performing credit scoring and credit rating to generate credit scoring result data; storing the credit scoring result data into a data storage device for a user to invoke; and invoking the credit scoring result data to calculate the retail asset risk according to a preset rule. The method and the system can be used for controlling the retail asset risk of the bank, can greatly reduce the dependence of bank credit personnel on the assessment criteria not objective and stable enough in the credit business, can effectively reduce randomness in decision and can reduce manual operation cost.
Description
Technical field
The present invention relates to field of computer technology, relate in particular to utilize communication network and computing machine to carry out the technical field of data processing, is a kind of computing method and system of retail asset risk specifically.
Background technology
Credit risk is meant that debtor or counterparty's fail to act contractual obligation or credit quality change, and influences the execution of contract, thereby brings the risk of loss for obligee or financial instrument possessor.Because to account for all professional proportions of bank higher for commercial bank's assets operation, so the basis risk that credit risk is a bank to be faced accurately discloses and the tolerance credit risk, to reducing the bank capital loss, it is vital improving operation result.The retail credit assets operation development of bank in the face of increasing fast will manage millions of personal credit clients and tens million of card holder's risk.Therefore, can the credit risk that estimate and predict bank's retail assets objective, accurately and efficiently be the key of the maintenance fast development of bank.
Bank has carried out the research of a lot of aspects to corporate client's asset risk evaluation in recent years, has got many achievements, but also stops the starting stage for the assessment of retail trade asset risk, can not reflect the loss size of retail assets expection effectively.
At present, bank mainly relies on people's experience to the evaluation of retail asset risk, and mainly by spreadsheet program instruments such as (as Micrsoft Excel and VBA macrocodes).There is following defective in this processing mode:
1. data acquisition relies on manually fully, and workload is huge, and data-handling efficiency is very low, limit by the spreadsheet program ability, is difficult to carry out the processing of big data quantity;
2. data are generally managed by the individual, and are destroyed easily, and safety of data does not ensure;
3. subjective judgement is occupied very big composition during to the new apply for loan of retail customer, to the evaluation neither one of its risk unified standard and computation model;
4. data are difficult to realize sharing, and evaluation result and experience are difficult to offer other and use use.
Summary of the invention
The present invention proposes in view of the above problems, and its purpose is, a kind of computing method and system of retail asset risk is provided, and is not accurate enough and not have a problem of seeking unity of standard to the evaluation of retail asset risk to solve.
The invention provides a kind of computing method of retail asset risk, this method comprises: gather source data, this source data comprises loan, credit card trade data, the customer capital data of gathering from each operation system of bank, client, account data during with the application of gathering in real time, and the client's reference data during from application that People's Bank of China's credit investigation system is gathered in real time from personal credit system and bank card system; Receive described source data and carry out the credit scoring processing and the credit rating processing, generate the credit scoring result data; Described credit scoring result data is stored in the data storage device for the user calls; Call described credit scoring result data to calculate the retail asset risk according to pre-defined rule.
The computing system of also a kind of retail asset risk of the present invention, this system comprises: data collector, be used to gather source data, this source data comprises loan, credit card trade data, the customer capital data of gathering from each operation system of bank, client, account data during with the application of gathering in real time, and the client's reference data during from application that People's Bank of China's credit investigation system is gathered in real time from personal credit system and bank card system; The credit scoring device is used to receive described source data and carries out the credit scoring processing and the credit rating processing, generates the credit scoring result data; Data storage device is used to store described credit scoring result data and supplies the user to call; The Risk Calculation device is used to call described credit scoring result data to calculate the retail asset risk according to pre-defined rule.
The beneficial effect of the embodiment of the invention is, the risk of may command bank retail assets, can significantly reduce the dependence of bank credit personnel to objective inadequately stable evaluation criteria in the credit operation, can effectively reduce the randomness in the decision, and can reduce the manually-operated cost, increase work efficiency, improve the service efficiency of bank capital.Can also reduce bank the human resources of case by case screening are carried out in application, this is particularly evident to the big credit product of portfolio.
Description of drawings
Accompanying drawing described herein is used to provide further understanding of the present invention, constitutes the application's a part, does not constitute limitation of the invention.In the accompanying drawings:
Shown in Figure 1A is the process flow diagram of computing method of the retail asset risk of the embodiment of the invention 1.
Shown in Figure 1B is the credit scoring processing flow chart of computing method of the retail asset risk of the embodiment of the invention 1.
Shown in Fig. 2 A is the structured flowchart of computing system of the retail asset risk of the embodiment of the invention 2.
Shown in Fig. 2 B is the structured flowchart of credit scoring device of computing system of the retail asset risk of the embodiment of the invention 2.
The structured flowchart of the computing system of the retail asset risk of the embodiment of the invention 3 that shown in Figure 3 is.
The process flow diagram of the computing method of the retail asset risk of the embodiment of the invention 3 that shown in Figure 4 is.
The process flow diagram of the step that the credit scoring of the embodiment of the invention 3 that shown in Figure 5 is is handled.
Shown in Figure 6 is the synoptic diagram of a typical segmentation scoring model of the embodiment of the invention 3.
Shown in Figure 7 is the credit rating processing flow chart of the embodiment of the invention 3.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer,, the present invention is described in further details below in conjunction with embodiment and accompanying drawing.At this, exemplary embodiment of the present invention and explanation thereof are used to explain the present invention, but not as a limitation of the invention.
Embodiment 1
Shown in Figure 1A is the process flow diagram of computing method of the retail asset risk of the embodiment of the invention 1.Shown in Figure 1A, the computing method of the retail asset risk of the embodiment of the invention 1 comprise:
S101: gather source data, this source data comprises loan, credit card trade data, the customer capital data of gathering from each operation system of bank, client, account data during with the application of gathering in real time, and the client's reference data during from application that People's Bank of China's credit investigation system is gathered in real time from personal credit system and bank card system;
S102: receive above-mentioned source data and carry out the credit scoring processing and the credit rating processing, generate the credit scoring result data;
S103: above-mentioned credit scoring result data is stored in the data storage device for the user calls;
S104: call above-mentioned credit scoring result data to calculate the retail asset risk according to pre-defined rule.
Shown in Figure 1B is the credit scoring processing flow chart of computing method of the retail asset risk of the embodiment of the invention 1.Shown in Figure 1B, in the embodiment of the invention 1, the step that above-mentioned credit scoring is handled can also comprise:
S105: obtain correlation parameter according to above-mentioned source data;
S106: according to above-mentioned correlation parameter, identical or similar target client is divided into same group behavior pattern, and construct decision tree automatically by segmentation different business kind, wherein the structure of decision tree adopts the level mode of going forward one by one, each condition of intermediate link is concrete division rule, finish node is a concrete scoring model, thus to not on the same group customers use different scoring models;
S107:, call corresponding scoring model and calculate above-mentioned credit scoring for each client or loan product.
Wherein, above-mentioned scoring model adopts following Logic Regression Models:
Y＝α+β
1Y
1+β
2Y
2+…+β
nY
n
Wherein, α is a constant, β
1β
2Be coefficient, Y1, Y2 are variablees, and Y is a result of calculation, α, β
1β
2, Y1, Y2 obtain from above-mentioned correlation parameter,
And, after calculating above-mentioned Y, carry out the mark that following The deformation calculation obtains client's credit scoring for Y:
SCORE＝round(1000/(1+e
(-Y)))。
By the embodiment of the invention 1, the risk of may command bank retail assets, can significantly reduce the dependence of bank credit personnel to objective inadequately stable evaluation criteria in the credit operation, can effectively reduce the randomness in the decision, and can reduce the manually-operated cost, increase work efficiency, improve the service efficiency of bank capital.Can also reduce bank the human resources of case by case screening are carried out in application, this is particularly evident to the big credit product of portfolio.
Embodiment 2
Shown in Fig. 2 A is the structured flowchart of computing system of the retail asset risk of the embodiment of the invention 2.Shown in Fig. 2 A, the computing system of the retail asset risk of the embodiment of the invention 2 comprises:
Credit scoring device 202 is used to receive above-mentioned source data and carries out the credit scoring processing and the credit rating processing, generates the credit scoring result data;
Data storage device 203 is used to store above-mentioned credit scoring result data and supplies the user to call;
Risk Calculation device 204 is used to call above-mentioned credit scoring result data to calculate the retail asset risk according to pre-defined rule.
Shown in Fig. 2 B is the structured flowchart of credit scoring device of computing system of the retail asset risk of the embodiment of the invention 2.Shown in Fig. 2 B, the credit scoring device of the computing system of the retail asset risk of the embodiment of the invention 2 comprises:
Decision tree generation unit 206, be used for according to above-mentioned correlation parameter, identical or similar target client is divided into same group behavior pattern, and construct decision tree automatically by segmentation different business kind, wherein the structure of decision tree adopts the level mode of going forward one by one, each condition of intermediate link is concrete division rule, and finish node is a concrete scoring model, thus to not on the same group customers use different scoring models;
Wherein, in the embodiment of the invention 2, above-mentioned scoring model adopts following Logic Regression Models:
Y＝α+β
1Y
1+β
2Y
2+…+β
nY
n
Wherein, α is a constant, β
1β
2Be coefficient, Y1, Y2 are variablees, and Y is a result of calculation, α, β
1β
2, Y1, Y2 obtain from above-mentioned correlation parameter,
And, after calculating above-mentioned Y, carry out the mark that following The deformation calculation obtains client's credit scoring for Y:
SCORE＝round(1000/(1+e
(-Y)))。
In addition, in the embodiment of the invention 2, above-mentioned Risk Calculation device comprises: credit rating unit (not shown), it is used for calculating Default Probability, default risk exposure, credit conversion coefficient, expected loss rate, credit risk capital requirement coefficient, unexpected loss, risk-weighted asset according to above-mentioned credit scoring result data.
Wherein, above-mentioned credit rating unit reads in a loan and an above-mentioned credit scoring result data, the retail assets is divided into the special assets pond by assets pond decision tree, and gives corresponding Default Probability.
By the invention described above embodiment 2, the risk of may command bank retail assets, can significantly reduce the dependence of bank credit personnel to objective inadequately stable evaluation criteria in the credit operation, can effectively reduce the randomness in the decision, and can reduce the manually-operated cost, increase work efficiency, improve the service efficiency of bank capital.Can also reduce bank the human resources of case by case screening are carried out in application, this is particularly evident to the big credit product of portfolio.
Embodiment 3
The embodiment of the invention 3 provides a kind of computing method and system of retail asset risk, it adopts logistic regression and sets up metering and the evaluation of the method realization of decision tree to bank's retail asset risk, can gather the source data of each operation system automatically, and these source datas are analyzed and handle, the evaluation result data storage of generation is in data storage device.
The structured flowchart of the computing system of the retail asset risk of the embodiment of the invention 3 that shown in Figure 3 is.As shown in Figure 3, the computing system of the retail asset risk of the embodiment of the invention 3 comprises: data collector 301, Risk Calculation device 302, data storage device 303.Between data collector 301 and the Risk Calculation device 302, can be connected by internal network between Risk Calculation device 302 and the data storage device 303.Internal network is the LAN (Local Area Network) of enterprise, can be Ethernet (Ethernet), also can be other LAN (Local Area Network), as Fiber Distributed Data Interface (FDDI), token ring (Token-Ring) etc.Can also the LAN (Local Area Network) of its each branch offices be connected to form bigger intranet (Intranet) by renting modes such as special line in addition.
Data collector 301 can be a minicomputer or PC server, this device is gathered source data (comprising the application information, assets information, reference information, Transaction Information of customer lending etc.) in the mode of data-interface from existing banking system, sends Risk Calculation device 302 to by internal network again.
Risk Calculation device 302 can be a minicomputer or PC server, is responsible for the source data that collects is estimated processing.
Data storage device 303 can be a minicomputer or PC server, the data of the evaluation result that storage Risk Calculation device 302 calculates.
The process flow diagram of the computing method of the retail asset risk of the embodiment of the invention 3 that shown in Figure 4 is.As shown in Figure 4, the computing method of the retail asset risk of the embodiment of the invention 3 comprise:
Step S401: data collector 301 beginning image data, from each operation system (individual loan, credit card etc.) of bank by data-interface collection loan, credit card trade data, customer capital data; Client, account data when personal credit system and bank card system are gathered application in real time by data-interface; Gather client's reference data in when application in real time by data-interface from People's Bank of China's credit investigation system; And send the source data that these collect to Risk Calculation device 302;
Step S402: the interface data that Risk Calculation device 302 is gathered data collector 301 measures processing, and the metering processing procedure comprises the credit scoring processing and credit rating is handled.
Step S403: Risk Calculation device 302 is stored in data storage device 303 with the result data of metering.
The process flow diagram of the step that the credit scoring of the embodiment of the invention 3 that shown in Figure 5 is is handled.As shown in Figure 5, the credit scoring of the embodiment of the invention 3 is handled and is comprised the steps:
Step S501: obtain correlation parameter.Obtain constant, coefficient and the variable of evaluation model, wherein constant, coefficient obtain from data storage device 303, variable comes from data collector 301, also needs to carry out in batches according to Different Rule processing such as missing values, exceptional value replacement, variable distortion for these variablees.
Step S502: construct decision tree automatically.According to parameters such as the scoring model that obtains, variable rules, construct decision tree automatically.The method of structure decision tree is: for identical product because different applicant's expression behaviour differences, identical or similar target client is divided into same group behavior pattern by the segmentation variable, to not on the same group customers use different scoring models, to distinguish different behavior patterns.The scoring model scheduling mode that different classs of business is corresponding different, therefore corresponding every kind of class of business needs independent generation model decision tree.The generation of decision tree adopts the level mode of going forward one by one to carry out, and each condition of intermediate link is the concrete division rule of segmentation variable, and finish node is to specify a terminal note and have a concrete scoring model.
The number of plies of decision tree and structure are provided with flexibly by constant, coefficient and the variable of evaluation model, and stronger extendability and reconstruct are arranged, and can adjust the structure of decision tree arbitrarily.
For example: a typical segmentation scoring model as shown in Figure 6, the client is subdivided into three layers, totally 5 colonies, these 5 colonies also claim final leaf node (node 1, node 5, node 6, node 7, node 8), to the corresponding respectively scoring model of each colony (final leaf node), calculate (as shown in table 1 below) by calling corresponding scoring model.
| Colony | Final leaf node | The scoring model that uses |
| 1 | Node 1 | Scoring model 1 |
| 2 | Node 5 | Scoring model 2 |
| 3 | Node 6 | Scoring model 3 |
| 4 | Node 7 | Scoring model 4 |
| 5 | Node 8 | Scoring model 5 |
Table 1
Step S503: call the scoring Model Calculation.
For each client or loan product, the decision tree that produces according to step S502, variable travels through under the father node, whether node is final leaf node below judging, until finding final leaf node,, call corresponding scoring model and calculate according to final leaf node.
The scoring model adopts Logic Regression Models.A typical Logic Regression Models is:
The Y=alpha+beta
1Y
1+ β
2Y
2+ ... + β
nY
n... formula 1
In the above-mentioned formula 1, α is a constant, β
1β
2Be coefficient, Y1Y2 is variable (as: age, an income etc.), and Y is a result of calculation.α, β
1β
2Can obtain from step 201, Y1 Y2 obtains from data collector 1.
After calculating above-mentioned Y, carry out the mark that following The deformation calculation just can obtain client's credit scoring for Y.
SCORE=round (1000/ (1+e
(-Y))) ... formula 2
Illustrate: certain card holder (Zhang San, age 25, income 2500,10,000 yuan of time deposits, there is not bad record), automatically match the node of decision tree according to client's parameter of obtaining (area, professional, the age bracket of credit card, whether record of bad behavior is arranged), and correspond to concrete model, obtain its coefficient: α=500, β according to model
1=0.1, Y
1Be that the variable income is 2500), substitution formula computing client scoring Y:
The Y=alpha+beta
1Y
1=500+0.1*2500=750 branch, drawing this client's credit scoring is 750 minutes.
Shown in Figure 7 is the credit rating processing flow chart of the embodiment of the invention 3.As shown in Figure 7, it is to carry out after aforementioned calculation goes out client's credit scoring that the credit rating of the embodiment of the invention 3 is handled again, and its step comprises:
Step S701: Default Probability PD metering.The metering of Default Probability (PD) be meant the metering each retail assets the grading time point Default Probability or estimate every retail application the application time point Default Probability.
Step a: read in a loan, credit card grading interface in the data collector, and above-mentioned risk score is handled good as calculated client's credit scoring;
Step b: the retail assets are divided into the special assets pond by assets pond decision tree;
Step c: in a single day every retail assets are included into the special assets pond, then give corresponding Default Probability (PD) assessed value.This assessed value is an empirical value, is that the mode by parameter is stored in the data storage device 3.
Step S702: promise breaking loss percentage LGD metering
The promise breaking loss percentage (LGD) metering be meant the metering each retail assets the grading time point the promise breaking loss percentage or estimate every retail application the application time point the promise breaking loss percentage, its metering method is identical with the PD metering method.
Step S703: default risk exposes the metering of (EAD) and credit conversion coefficient (CCF).The metering that default risk exposes (EAD) is meant that estimating each retail assets exposes or estimate that every retail application exposes in the default risk of application time point in the default risk of grading time point.Its concrete calculation process is as follows:
Step a: read in the borrowing of new application in the data collector and storage, credit card grading interface, and risk score handle in good as calculated credit scoring;
Step b: differentiate the feature of risk variable successively: " class of assets ", " promise breaking state ", select the metering formula to calculate then.For QR class, non-promise breaking assets, need these retail assets of metering CCF value earlier before measuring;
Step c: metering EAD.
The loan balance=refer to of promise breaking time point/grading time point break a contract the loan principal of time point/grading time point+should go back interest+fee payable with.
Overdraw remaining sum=overdraw capital+should go back interest+fee payable.
The special processing of overdraw remaining sum: if do not overdraw or have money on deposit, the remaining sum of then overdrawing is 0.
The promise breaking time point/grading time point loan principal, should go back interest, fee payable; The overdraw capital, should go back interest, fee payable and all from data collector 1, obtain.
Concrete computing formula is as follows:
(1) EAD of storage retail assets metering formula is as shown in table 2 below:
Table 2
(2) the EAD metering of application retail assets is as shown in table 3 below:
| Not promise breaking (NID) | |
| MR | The accrediting amount of grading time point loan application |
| OR | The accrediting amount of grading time point loan application |
| QR | CCF *The initial credit amount that the grading time point is judged |
Table 3
The metering of credit conversion coefficient (CCF) is meant that metering may increase the risk exposure of use over a period to come during to customer default from the grading time point.The method of CCF metering is with PD/LGD metering method unanimity.
Step S704: expected loss rate (EL Ratio) and expected loss (EL) metering.
The metering of loss is meant at the grading time point, for arbitrary retail application business or storage retail assets, according to its Default Probability, promise breaking loss percentage and default risk exposure, the isoparametric estimation of credit conversion coefficient, measure the ratio and the degree of this assets loss.
Expected loss rate ELR=PD*LGD
Expected loss EL=PD*LGD*EAD
PD, LGD, EAD obtain from the step of front.
Step S705: the credit risk capital requires coefficient (K) and unexpected loss (UL) metering
(1), the credit risk capital requires the metering of coefficient (K)
It in fact also is non-expected loss rate that the credit risk capital requires coefficient (K).
For non-promise breaking assets:
Wherein, R is the assets correlativity, and its size is decided according to the class of assets in the interface, and the metering formula is:
Wherein, PD, LGD are meant the risk parameter estimation of grading time point assets, obtain from preceding step.
For the promise breaking assets:
K=max (0, (PLGD-BEEL)) ... formula 4
Wherein PLGD, BEEL grading time point obtains from the step of front the estimation of promise breaking loss percentage.
(2), the metering of unexpected loss
It is unexpected that loss can be regarded as the shared economic capital gold of single assets again.The metering formula is:
UL=K*EAD=EC...... formula 5
Wherein K requires coefficient for the credit risk capital.
Step S706: the metering of risk-weighted asset (RWA).
Risk-weighted asset for capital controls service, metering formula is:
RWA=K * 12.5 * EAD...... formula 6
Wherein, K requires coefficient for the credit risk capital; The EAD that EAD determines for the grading time point.
The embodiment of the invention 3 has overcome present bank the retail asset risk has been quantized and the main human factor that relies on of evaluation, can not realize difficulty objective, accurate, efficient evaluations, a kind of objective, accurate, method that treatment effeciency is high estimates bank's retail asset risk is provided.
Beneficial effect of the present invention is:
1, improves data-handling efficiency and security, realize carrying out the processing of big data quantity.
2, realize beginning performance with regard to the dynamic tracking client from client's apply for loan, the risk of control bank retail assets, and improve income.Begin just to follow the tracks of client's performance from the client at commercial bank's apply for loan, prediction customer default risk is taken precautions against top-tier customer and is run off, carries out targetedly marketing strategy, supports for the market competitiveness that further strengthens commercial bank's products ﹠ services provides stable.
3, the dependence of bank credit personnel can be significantly reduced, the randomness in the decision can be effectively reduced objective inadequately stable evaluation criteria in the credit operation.Can use the scoring model to come unified quantitative evaluation credit risk, to different regions, different business kind, different period, the performed credit management policy of different credit personnel is weighed and relatively.
5, data realize sharing, and evaluation result and experience are difficult to offer other and use use.
6, can reduce the manually-operated cost, increase work efficiency, improve the service efficiency of bank capital.Can also reduce bank the human resources of case by case screening are carried out in application, this is particularly evident to the big credit product of portfolio.
Above-described embodiment; purpose of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the above only is the specific embodiment of the present invention; and be not intended to limit the scope of the invention; within the spirit and principles in the present invention all, any modification of being made, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (10)
1. the computing method of a retail asset risk is characterized in that, this method comprises:
Gather source data, this source data comprises loan, credit card trade data, the customer capital data of gathering from each operation system of bank, client, account data during with the application of gathering in real time, and the client's reference data during from application that People's Bank of China's credit investigation system is gathered in real time from personal credit system and bank card system;
Receive described source data and carry out the credit scoring processing and the credit rating processing, generate the credit scoring result data;
Described credit scoring result data is stored in the data storage device for the user calls;
Call described credit scoring result data to calculate the retail asset risk according to pre-defined rule.
2. the computing method of retail asset risk according to claim 1 is characterized in that, the step that described credit scoring is handled comprises:
Obtain correlation parameter according to described source data;
According to described correlation parameter, identical or similar target client is divided into same group behavior pattern, and construct decision tree automatically by segmentation different business kind, wherein the structure of decision tree adopts the level mode of going forward one by one, each condition of intermediate link is concrete division rule, finish node is a concrete scoring model, thus to not on the same group customers use different scoring models;
For each client or loan product, call corresponding scoring model and calculate described credit scoring.
3. the computing method of retail asset risk according to claim 2 is characterized in that,
Described scoring model adopts following Logic Regression Models:
Y＝α+β
1Y
1+β
2Y
2+…+β
nY
n
Wherein, α is a constant, β
1β
2Be coefficient, Y1, Y2 are variablees, and Y is a result of calculation, α, β
1β
2, Y1, Y2 obtain from described correlation parameter,
And, after calculating described Y, carry out the mark that following The deformation calculation obtains client's credit scoring for Y:
SCORE＝round(1000/(1+e
(-Y)))。
4. the computing method of retail asset risk according to claim 2 is characterized in that, the step of described calculating retail asset risk also comprises:
Calculate Default Probability, default risk exposure, credit conversion coefficient, expected loss rate, credit risk capital requirement coefficient, unexpected loss, risk-weighted asset according to described credit scoring result data.
5. the computing method of retail asset risk according to claim 4 is characterized in that,
The step of described calculating Default Probability comprises: read in a loan and a described credit scoring result data, the retail assets are divided into the special assets pond by assets pond decision tree, and give corresponding Default Probability.
6. the computing system of a retail asset risk is characterized in that, this system comprises:
Data collector, be used to gather source data, this source data comprises loan, credit card trade data, the customer capital data of gathering from each operation system of bank, client, account data during with the application of gathering in real time, and the client's reference data during from application that People's Bank of China's credit investigation system is gathered in real time from personal credit system and bank card system;
The credit scoring device is used to receive described source data and carries out the credit scoring processing and the credit rating processing, generates the credit scoring result data;
Data storage device is used to store described credit scoring result data and supplies the user to call;
The Risk Calculation device is used to call described credit scoring result data to calculate the retail asset risk according to pre-defined rule.
7. the computing system of retail asset risk according to claim 6 is characterized in that, described credit scoring device comprises:
Parameter acquiring unit is used for obtaining correlation parameter according to described source data;
The decision tree generation unit, be used for according to described correlation parameter, identical or similar target client is divided into same group behavior pattern, and construct decision tree automatically by segmentation different business kind, wherein the structure of decision tree adopts the level mode of going forward one by one, each condition of intermediate link is concrete division rule, and finish node is a concrete scoring model, thus to not on the same group customers use different scoring models;
Computing unit is used for calling corresponding scoring model and calculating described credit scoring for each client or loan product.
8. the computing system of retail asset risk according to claim 7 is characterized in that,
Described scoring model adopts following Logic Regression Models:
Y＝α+β
1Y
1+β
2Y
2+…+β
nY
n
Wherein, α is a constant, β
1β
2Be coefficient, Y1, Y2 are variablees, and Y is a result of calculation, α, β
1β
2, Y1, Y2 obtain from described correlation parameter,
And, after calculating described Y, carry out the mark that following The deformation calculation obtains client's credit scoring for Y:
SCORE＝round(1000/(1+e
(-Y)))。
9. the computing system of retail asset risk according to claim 7, it is characterized in that, described Risk Calculation device comprises: the credit rating unit is used for calculating Default Probability, default risk exposure, credit conversion coefficient, expected loss rate, credit risk capital requirement coefficient, unexpected loss, risk-weighted asset according to described credit scoring result data.
10. the computing system of retail asset risk according to claim 9 is characterized in that,
Described credit rating unit reads in a loan and a described credit scoring result data, the retail assets is divided into the special assets pond by assets pond decision tree, and gives corresponding Default Probability.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100909367A CN101996381A (en) | 2009-08-14 | 2009-08-14 | Method and system for calculating retail asset risk |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100909367A CN101996381A (en) | 2009-08-14 | 2009-08-14 | Method and system for calculating retail asset risk |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101996381A true CN101996381A (en) | 2011-03-30 |
Family
ID=43786510
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009100909367A Pending CN101996381A (en) | 2009-08-14 | 2009-08-14 | Method and system for calculating retail asset risk |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101996381A (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104182897A (en) * | 2013-05-28 | 2014-12-03 | 中国农业银行股份有限公司天津市分行 | Capital return pricing management system |
| CN104463673A (en) * | 2014-12-22 | 2015-03-25 | 中国科学技术大学苏州研究院 | P2P network credit risk assessment model based on support vector machine |
| CN105590216A (en) * | 2015-11-18 | 2016-05-18 | 中国银联股份有限公司 | Method and system of real-time monitoring of transaction risk |
| CN105894372A (en) * | 2016-06-13 | 2016-08-24 | 腾讯科技（深圳）有限公司 | Method and device for predicting group credit |
| CN104063601B (en) * | 2014-06-26 | 2017-12-12 | 浙江德清沃尔赴金融数据处理有限公司 | The monitoring method and system calculated based on small micro- loan assets pond loss late |
| WO2017215403A1 (en) * | 2016-06-12 | 2017-12-21 | 腾讯科技（深圳）有限公司 | Method and apparatus for assessing user credit, and storage medium |
| CN107633455A (en) * | 2017-09-04 | 2018-01-26 | 深圳市华傲数据技术有限公司 | Credit estimation method and device based on data model |
| CN107808339A (en) * | 2017-09-07 | 2018-03-16 | 广西钱盆科技股份有限公司 | A kind of data acquisition and analysis system |
| CN108399543A (en) * | 2018-01-23 | 2018-08-14 | 阿里巴巴集团控股有限公司 | Binding method, method for evaluating trust, device and the electronic equipment of Payment Card |
| CN108460643A (en) * | 2017-02-20 | 2018-08-28 | 平安科技（深圳）有限公司 | Customer information processing method and processing device |
| CN108876076A (en) * | 2017-05-09 | 2018-11-23 | 中国移动通信集团广东有限公司 | The personal credit methods of marking and device of data based on instruction |
| CN109345381A (en) * | 2018-12-19 | 2019-02-15 | 重庆誉存大数据科技有限公司 | A kind of Risk Identification Method and system |
| CN109345261A (en) * | 2018-08-21 | 2019-02-15 | 上海淇毓信息科技有限公司 | A kind of credit cost automatic evaluation system |
| CN109840676A (en) * | 2018-12-13 | 2019-06-04 | 平安科技（深圳）有限公司 | Air control method, apparatus, computer equipment and storage medium based on big data |
| CN110135973A (en) * | 2019-04-23 | 2019-08-16 | 北京淇瑀信息科技有限公司 | A kind of intelligent credit method based on IM and intelligent credit device |
| CN110851540A (en) * | 2019-10-28 | 2020-02-28 | 天津大学 | Financial service map-based commercial bank customer loss early warning method |
| CN111507821A (en) * | 2020-04-09 | 2020-08-07 | 苏宁消费金融有限公司 | Asset pool wind control method and system based on rating and memorability strategies |
| CN112418580A (en) * | 2019-08-22 | 2021-02-26 | 上海哔哩哔哩科技有限公司 | Risk control method, computer equipment and readable storage medium |
| CN113191873A (en) * | 2021-04-23 | 2021-07-30 | 武汉赢联数据技术股份有限公司 | Bank credit card customer risk early warning system based on cloud data management |
-
2009
- 2009-08-14 CN CN2009100909367A patent/CN101996381A/en active Pending
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104182897A (en) * | 2013-05-28 | 2014-12-03 | 中国农业银行股份有限公司天津市分行 | Capital return pricing management system |
| CN104063601B (en) * | 2014-06-26 | 2017-12-12 | 浙江德清沃尔赴金融数据处理有限公司 | The monitoring method and system calculated based on small micro- loan assets pond loss late |
| CN104463673A (en) * | 2014-12-22 | 2015-03-25 | 中国科学技术大学苏州研究院 | P2P network credit risk assessment model based on support vector machine |
| CN105590216A (en) * | 2015-11-18 | 2016-05-18 | 中国银联股份有限公司 | Method and system of real-time monitoring of transaction risk |
| WO2017215403A1 (en) * | 2016-06-12 | 2017-12-21 | 腾讯科技（深圳）有限公司 | Method and apparatus for assessing user credit, and storage medium |
| CN105894372A (en) * | 2016-06-13 | 2016-08-24 | 腾讯科技（深圳）有限公司 | Method and device for predicting group credit |
| CN108460643A (en) * | 2017-02-20 | 2018-08-28 | 平安科技（深圳）有限公司 | Customer information processing method and processing device |
| CN108460643B (en) * | 2017-02-20 | 2021-05-11 | 平安科技（深圳）有限公司 | Client information processing method and device |
| CN108876076A (en) * | 2017-05-09 | 2018-11-23 | 中国移动通信集团广东有限公司 | The personal credit methods of marking and device of data based on instruction |
| CN107633455A (en) * | 2017-09-04 | 2018-01-26 | 深圳市华傲数据技术有限公司 | Credit estimation method and device based on data model |
| CN107808339A (en) * | 2017-09-07 | 2018-03-16 | 广西钱盆科技股份有限公司 | A kind of data acquisition and analysis system |
| CN108399543A (en) * | 2018-01-23 | 2018-08-14 | 阿里巴巴集团控股有限公司 | Binding method, method for evaluating trust, device and the electronic equipment of Payment Card |
| US10902415B2 (en) | 2018-01-23 | 2021-01-26 | Advanced New Technologies Co., Ltd. | Payment card binding method, trust evaluation method, apparatus, and electronic device |
| CN109345261A (en) * | 2018-08-21 | 2019-02-15 | 上海淇毓信息科技有限公司 | A kind of credit cost automatic evaluation system |
| CN109840676A (en) * | 2018-12-13 | 2019-06-04 | 平安科技（深圳）有限公司 | Air control method, apparatus, computer equipment and storage medium based on big data |
| CN109840676B (en) * | 2018-12-13 | 2023-09-15 | 平安科技（深圳）有限公司 | Big data-based wind control method and device, computer equipment and storage medium |
| CN109345381A (en) * | 2018-12-19 | 2019-02-15 | 重庆誉存大数据科技有限公司 | A kind of Risk Identification Method and system |
| CN110135973A (en) * | 2019-04-23 | 2019-08-16 | 北京淇瑀信息科技有限公司 | A kind of intelligent credit method based on IM and intelligent credit device |
| CN112418580A (en) * | 2019-08-22 | 2021-02-26 | 上海哔哩哔哩科技有限公司 | Risk control method, computer equipment and readable storage medium |
| CN110851540A (en) * | 2019-10-28 | 2020-02-28 | 天津大学 | Financial service map-based commercial bank customer loss early warning method |
| CN111507821A (en) * | 2020-04-09 | 2020-08-07 | 苏宁消费金融有限公司 | Asset pool wind control method and system based on rating and memorability strategies |
| CN111507821B (en) * | 2020-04-09 | 2022-08-02 | 苏宁消费金融有限公司 | Asset pool wind control method and system based on rating and memorability strategies |
| CN113191873A (en) * | 2021-04-23 | 2021-07-30 | 武汉赢联数据技术股份有限公司 | Bank credit card customer risk early warning system based on cloud data management |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101996381A (en) | Method and system for calculating retail asset risk | |
| CN107103548A (en) | The monitoring method and system and risk monitoring and control method and system of network behavior data | |
| CN108734565A (en) | A kind of reference point is real-time to adjust processing method, device and processing server | |
| CN101236638A (en) | Web based bank card risk monitoring method and system | |
| CN109074567A (en) | Fraction fund is transferred to savings system, program and method | |
| Utkulu et al. | Trade and competitiveness between Turkey and the EU: time series evidence | |
| KR20170049328A (en) | Apparatus and method for evaluating a value of technology or intellectual property | |
| Xuan | Regression analysis of supply chain financial risk based on machine learning and fuzzy decision model | |
| CN112132438A (en) | Investment and financing docking and equity value evaluation system for regional equity market | |
| Emadi et al. | A structural estimation approach to study agent attrition | |
| Gadatsch et al. | German and the rest of euro area fiscal policy during the crisis | |
| CN106803166A (en) | A kind of financial data processing method and system | |
| Bataev | Financial technology: Efficiency evaluation of challenger banks | |
| CN110322347A (en) | A kind of shot and long term strategy multiple-factor quantization capitalized method and device | |
| Mačerinskiene et al. | The evaluation model of a commercial bank loan portfolio | |
| Rachmadi et al. | The Credit Restructuring Phenomenon of The MSMEs and its Effect on Banking Financial Performance During The Pandemic of Covid-19 | |
| Leszczyńska-Paczesna | Sectoral price stickiness and inflation persistence in Poland: a two-sector DSGE approach | |
| CN110162572A (en) | Policy execution method, server and computer storage medium | |
| CN109360085A (en) | A kind of bank client responsible investigation method and system | |
| Zhu | Research on Risk Management and Countermeasures of Inclusive Finance——From the Perspective of Comprehensive Risk Management of Credit Insurance | |
| TWI720638B (en) | Deposit interest rate bargaining adjustment system and method thereof | |
| Kovalova et al. | Innovate forms of project financing under the conditions of globalization | |
| Olayiwola et al. | Trade liberalization and employment performance of textile and clothing industry in Tanzania | |
| Jeong et al. | Does borrowers' information renewal change lenders' decision in P2P lending? an empirical investigation | |
| CN101162522A (en) | System and method for metering bank economy capital |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Open date: 20110330 |