CN108492001A - A method of being used for guaranteed loan network risk management - Google Patents
A method of being used for guaranteed loan network risk management Download PDFInfo
- Publication number
- CN108492001A CN108492001A CN201810150018.8A CN201810150018A CN108492001A CN 108492001 A CN108492001 A CN 108492001A CN 201810150018 A CN201810150018 A CN 201810150018A CN 108492001 A CN108492001 A CN 108492001A
- Authority
- CN
- China
- Prior art keywords
- loan
- risk management
- model
- window
- assessment models
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
Abstract
The present invention relates to a kind of methods for guaranteed loan network risk management, include the following steps:Step 1 carries out selective use by data acquisition module to all client's essential informations in guaranteed loan network;Step 2 acquires information to step 1 by characteristic vector pickup module and carries out five kinds of behaviors divisions;Step 3 divides five kinds of behavior structure promise breaking assessment models according to step 2;Step 4 acquires assessment models of breaking a contract in a certain client's essential information input step three in collateral warranty loan network again and analyzes;Step 5 checks that the assessment models carry out Anticipatory breach judging result to a certain client by visual windows module, and this method can improve the accuracy of bank lending risks promise breaking assessment.
Description
Technical field
The present invention relates to banking management system technical fields, and in particular to a kind of for guaranteed loan network risk management
Method.
Background technology
Risk management is the core of all financial business, is always the emphasis of local government and bank about financial security
Attentinal contents.So-called enterprise security network (also known as guarantee circle) refers to the net being interconnected to form by assuring contract between enterprise
Network structure organization.It is difficult derived from government's reply Financing that guarantee is enclosed, this worldwide restricts economic development
Problem.Usual Corporate finance is mainly provided a loan from bank in addition to listing.However economic field it is more active and more lack money
The medium-sized and small enterprises of gold, often not only having lacked profit scale can not list, but also require qualification because not meeting bank batch loan, it is difficult to obtain
Bank loan.In order to promote medium-sized and small enterprises to develop, European and American developed countries, as the U.S. and British government are set up by government's background
Loan guarantee mechanism, to help medium-sized and small enterprises to obtain fund from business bank.However in East Asia, mainly in China and South Korea, silver
It is that it is assured to reduce risk that row, which usually requires that loan enterprises voluntarily find third party enterprise,.
Currently, loans on bank guarantee flow:The medium-sized and small enterprises (borrowing enterprise) of qualification are not met in order to be borrowed from bank
Money, it is necessary first to and several third company sign contract of guaranty, then sign loan contract with bank, just will receive silver later
Row fund and periodically repaying.
In fact, guaranteed loan has become one of the main channel of Small and Medium Enterprises in China financing.In more and more
Small business is associated by assuring mutually, and guarantee relationship also becomes to become increasingly complex, and ultimately generates the loan of several complexity
Assure network.This brings unprecedented challenge to the risk management of bank.Complicated connected guarantee loan present in reality, for
The sound development of national economy is a double-edged sword.In economic situation upward period, guaranteed loan, which can meet, does not meet qualification
Medium-sized and small enterprises financing needs, promote the development of private economy.But in economic descending phase, complex network is in Shi Huan enterprises
While default risk, it is also possible to lead to break a contract on a large scale infection and propagation phenomenon.If single event of default passes through network
It propagates, can lead to tide of breaking a contract on a large scale, this can be to regional industry, bank capital, and social security etc. causes a significant threat.
The great systemic financial security hidden danger that enterprise connected guarantee loan is brought includes:Fund is extracted, enterprise's excessive financing is very
Extremely it is easy forming region, system risk of industry etc..However traditional loans on bank guarantee security strategies, to increasingly complicated
Guarantee network risk management problem faces enormous challenge:Existing loan examines that monitoring, anti-fraud and decision-making mechanism lag behind
The market demand, the main large enterprise of current bank loan valuation method and design, credit evaluation usually mainly consider enterprise wealth
Business information, it is difficult to consider dependence of the borrower in increasingly complicated guarantee network.
Invention content
It is an object of the invention to overcome defect existing for above-mentioned background technology, provide a kind of for bank lending risks pipe
The system for managing assessment, this method can improve the accuracy of bank lending risks promise breaking assessment.
Technical scheme of the present invention:
A method of it being used for guaranteed loan network risk management, is included the following steps:
Step 1 carries out selectivity to all client's essential informations in guaranteed loan network by data acquisition module and adopts
With;
Step 2 acquires information to step 1 by characteristic vector pickup module and carries out five kinds of behaviors divisions;
Step 3 divides five kinds of behavior structure promise breaking assessment models according to step 2;
Step 4 acquires promise breaking assessment mould in a certain client's essential information input step three again in collateral warranty loan network
Type is analyzed;
Step 5 checks that the assessment models carry out Anticipatory breach to a certain client and judge knot by visual windows module
Fruit.
The step 2 characteristic vector pickup module divides five kinds of behaviors:Basic situation, behavior of credit, active loan
Money, network structure and community's behavior.
Assessment models of breaking a contract in step 3 foundation is to carry out in accordance with the following steps:
Step 1 can be expressed as using the tree aggregation model of the K function prediction summed it up output:
Wherein fkFor kth decision tree, XiFor trained feature,For prediction result.
Step 2, the parameter for finding tree-model are converted into minimum objective function problem, and loss function can be defined as:
Wherein, formula 3 indicates training error (training error), for weighing the prediction of model on the training data
Ability;Formula 4 indicates regular terms, is used for the complexity of Controlling model, by increasing regularization term, encourages simpler mould
Type, and prevent over-fitting;Wherein γ, λ are adjustable weight parameter, and T is the leaf node number of tree-model;ω is leaf node power
Weight.The visual windows module includes training window, observation window, prediction window and evaluation window.
Compared with prior art, the present invention has the advantage that:
1, the present invention propose financial circles guarantee network assessing credit risks mixing express and combine supervised learning method and
Sliding time window realizes the prediction to enterprise's violation of agreement under guarantee network.Experiment shows that enterprise attributes, network structure are gone through
History repaying information, community's promise breaking information etc. have stronger correlation with promise breaking, as guarantee network becomes more and more complicated, network
Structure-related characteristic also becomes more and more important in violation correction.
2, present invention could apply to the risk assessment to enterprise of bank loan early period, help more accurately to find
The venture business being related in complicated guarantee network, reduces the risk of loss of bank loan.
Description of the drawings
Fig. 1 is a kind of method flow diagram for guaranteed loan network risk management of the invention.
Fig. 2 is to generate violation correction result figure using a kind of method for guaranteed loan network risk management of the present invention.
Specific implementation mode
Below by specific embodiments and the drawings, the present invention is further illustrated.The embodiment of the present invention is in order to more
So that those skilled in the art is more fully understood the present invention well, any limitation is not made to the present invention.
As shown in Figure 1, a kind of method for guaranteed loan network risk management of the invention, includes the following steps:
Step 1 101 carries out selectivity by data acquisition module to all client's essential informations in guaranteed loan network
Using;
Bank needs to collect the fine granularity information as much as possible about enterprise's loan repayment capacity.To ensure privacy, original number
Customer name in has been encrypted and has replaced with individual ID.These information are broadly divided into four classes:Transaction Information, Ke Huxin
Breath, the assets informations such as mortgage state and history loan approval banker's record etc..Maximally related with guaranteed loan is eight databases
Table:Customer data (Customer Profile), loan account information (Loan Account Info), refund state
(Repayment Status), guarantee situation (Guarantee Profile), client's prestige (Customer Credit), loan
Contract (Loan Contract), guarantee relationship (Guarantee Relationship), contract of guaranty (Guarantee
Contyract), promise breaking state (Default Status).By by the processing of these database table connection types (join), obtaining
Company ID and the relevant record of loan agreement, and it is based on guarantee relationship, structure guarantee network.
Step 2 102 acquires information to step 1 by characteristic vector pickup module and carries out five kinds of behaviors divisions;
1, basic situation (Basic Profile) refers to basic register of company's information, reflects personage, capital, guarantee,
Firms profitability, condition and stability.We are with business nature, registered capital, scope of the enterprise, headcount etc. for company's base
This situation.Most of banks require enterprise to update essential information in business loan application, we select to make using up-to-date information
For the basic situation of loan.
2, behavior of credit (Credit Behavior) refers to the credit history relevant information before enlivening loan agreement, including
Credit record, promise breaking record, the promise breaking amount of money/loan ceiling and loan value, loan ceiling, average rate of violation.
3, the loan agreement during actively loan (Active Loan) is execution, including loan limit is enlivened, span of providing a loan
With repaying mode etc..
4, network structure (Network Structure) refers to network characterization such as centrad etc., it is notable that enterprise
Basic financial information may be not exclusively credible, because medium-sized and small enterprises may be out-of-date to bank's offer or even fabricate letter
Breath.However, guarantee network is reliable information, because bank can build it from the record system of oneself.
5, as previously shown, promise breaking has group phenomenon, and breaking a contract can for community's behavior (Community Behavior)
It can be propagated as infectious disease in community.Therefore, the average rate of violation of independent community, used also as community's behavioural characteristic.
Step 3 103 divides five kinds of behavior structure promise breaking assessment models according to step 2;It breaks a contract and assesses in the step 3
Model foundation is to carry out in accordance with the following steps:
Step 1 can be expressed as using the tree aggregation model of the K function prediction summed it up output:
Wherein fkFor kth decision tree, XiFor trained feature,For prediction result.
Step 2, the parameter for finding tree-model are converted into minimum objective function problem, and loss function can be defined as:
Wherein, formula 3 indicates training error (training error), for weighing the prediction of model on the training data
Ability;Formula 4 indicates regular terms, is used for the complexity of Controlling model, by increasing regularization term, encourages simpler mould
Type, and prevent over-fitting;Wherein γ, λ are adjustable weight parameter, and T is the leaf node number of tree-model;ω is leaf node power
Weight.
Step 4 104 is acquired to break a contract in a certain client's essential information input step three and be commented in collateral warranty loan network again
Estimate model to be analyzed;
Step 5 105 checks that the assessment models carry out Anticipatory breach judgement to a certain client by visual windows module
As a result.
Risk profile based on sliding time window.First, the loan documentation in data warehouse is extracted and stored in client
Data management system (CDM), we introduce the mode of sliding time window, extraction feature as prediction algorithm training data with
Test data.Why sliding time window is used, is because of economic environment, enterprise operation is dynamic change, in order to make spy
Sign has stronger ability to express, we introduce sliding time window and extract enterprise characteristic in window.Here, time window
It is defined as:
Training window (training window):Within the time period, extraction enterprise composite character;
Observation window (observation window):Within the time period, label whether corresponding enterprise's promise breaking is obtained
Information is used for training pattern;
Prediction window (Predicting window):By trained assessment models for predicting enterprise in this time window
Whether break a contract in mouthful;
Evaluation window (Evaluation window):Enterprise's promise breaking state within the time period be used to verify prediction
Window prediction result.
In practice, it is analyzed according to business demand, we set length of window as three months.Fig. 2 gives sliding window
Schematic diagram.By taking left slide window as an example, we extract first quarter company information in 2013 as training data, and with this
A little enterprises the second quarter refund state as label.Whether the model trained is used for the prediction enterprise of the second quarter in 2013
Promise breaking, and assessment algorithm effect is come using practical refund situation with the third season.
It should be understood that embodiment and example discussed herein simply to illustrate that, to those skilled in the art
For, it can be improved or converted, and all these modifications and variations should all belong to the protection of appended claims of the present invention
Range.
Claims (4)
1. a kind of method for guaranteed loan network risk management, which is characterized in that include the following steps:
Step 1 carries out selective use by data acquisition module to all client's essential informations in guaranteed loan network;
Step 2 acquires information to step 1 by characteristic vector pickup module and carries out five kinds of behaviors divisions;
Step 3 divides five kinds of behavior structure promise breaking assessment models according to step 2;
Step 4, again collateral warranty loan network in acquire in a certain client's essential information input step three break a contract assessment models into
Row analysis;
Step 5 checks that the assessment models carry out Anticipatory breach judging result to a certain client by visual windows module.
2. a kind of method for guaranteed loan network risk management according to claim 1, it is characterised in that:The step
Rapid two characteristic vector pickups module divides five kinds of behaviors:Basic situation, behavior of credit, active loan, network structure and
Community's behavior.
3. a kind of method for guaranteed loan network risk management according to claim 1, it is characterised in that:The step
Assessment models of breaking a contract in rapid three foundation is to carry out in accordance with the following steps:
Step 1 can be expressed as using the tree aggregation model of the K function prediction summed it up output:
Wherein fkFor kth decision tree, XiFor trained feature,For prediction result.
Step 2, the parameter for finding tree-model are converted into minimum objective function problem, and loss function can be defined as:
Wherein, formula 3 indicates training error (training error), for weighing the prediction energy of model on the training data
Power;Formula 4 indicates regular terms, is used for the complexity of Controlling model, by increase regularization term, encourages simpler model,
And prevent over-fitting;Wherein γ, λ are adjustable weight parameter, and T is the leaf node number of tree-model;ω is leaf node weight.
4. a kind of method for guaranteed loan network risk management according to claim 1, it is characterised in that:It is described can
Viewing window module includes training window, observation window, prediction window and evaluation window.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810150018.8A CN108492001A (en) | 2018-02-13 | 2018-02-13 | A method of being used for guaranteed loan network risk management |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810150018.8A CN108492001A (en) | 2018-02-13 | 2018-02-13 | A method of being used for guaranteed loan network risk management |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108492001A true CN108492001A (en) | 2018-09-04 |
Family
ID=63340517
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810150018.8A Pending CN108492001A (en) | 2018-02-13 | 2018-02-13 | A method of being used for guaranteed loan network risk management |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108492001A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109255506A (en) * | 2018-11-22 | 2019-01-22 | 重庆邮电大学 | A kind of internet finance user's overdue loan prediction technique based on big data |
CN109508864A (en) * | 2018-10-19 | 2019-03-22 | 南京理工大学 | A kind of method for building up of enterprise's default risk model based on xgboost |
CN109816245A (en) * | 2019-01-25 | 2019-05-28 | 北京海致星图科技有限公司 | For conducting assessment system and method to the risk of public credit customer risk early warning |
CN109961362A (en) * | 2019-02-19 | 2019-07-02 | 合肥工业大学 | P2P platform credit risk dynamic evaluation method and system |
CN111861707A (en) * | 2020-07-16 | 2020-10-30 | 天津大学 | Quantification and visual processing method for infection risk of guarantee network |
CN112308294A (en) * | 2020-10-10 | 2021-02-02 | 北京贝壳时代网络科技有限公司 | Default probability prediction method and device |
CN113240509A (en) * | 2021-05-18 | 2021-08-10 | 重庆邮电大学 | Loan risk assessment method based on multi-source data federal learning |
CN114155093A (en) * | 2022-02-08 | 2022-03-08 | 一方函互联网有限公司 | Block chain-based electronic insurance full-flow management and risk management and control system |
-
2018
- 2018-02-13 CN CN201810150018.8A patent/CN108492001A/en active Pending
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109508864A (en) * | 2018-10-19 | 2019-03-22 | 南京理工大学 | A kind of method for building up of enterprise's default risk model based on xgboost |
CN109508864B (en) * | 2018-10-19 | 2022-08-05 | 南京理工大学 | Method for establishing enterprise default risk model based on xgboost |
CN109255506A (en) * | 2018-11-22 | 2019-01-22 | 重庆邮电大学 | A kind of internet finance user's overdue loan prediction technique based on big data |
CN109255506B (en) * | 2018-11-22 | 2022-05-03 | 重庆邮电大学 | Internet financial user loan overdue prediction method based on big data |
CN109816245A (en) * | 2019-01-25 | 2019-05-28 | 北京海致星图科技有限公司 | For conducting assessment system and method to the risk of public credit customer risk early warning |
CN109961362A (en) * | 2019-02-19 | 2019-07-02 | 合肥工业大学 | P2P platform credit risk dynamic evaluation method and system |
CN111861707A (en) * | 2020-07-16 | 2020-10-30 | 天津大学 | Quantification and visual processing method for infection risk of guarantee network |
CN112308294A (en) * | 2020-10-10 | 2021-02-02 | 北京贝壳时代网络科技有限公司 | Default probability prediction method and device |
CN113240509A (en) * | 2021-05-18 | 2021-08-10 | 重庆邮电大学 | Loan risk assessment method based on multi-source data federal learning |
CN113240509B (en) * | 2021-05-18 | 2022-04-22 | 重庆邮电大学 | Loan risk assessment method based on multi-source data federal learning |
CN114155093A (en) * | 2022-02-08 | 2022-03-08 | 一方函互联网有限公司 | Block chain-based electronic insurance full-flow management and risk management and control system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Ma et al. | Financial credit risk prediction in internet finance driven by machine learning | |
CN108492001A (en) | A method of being used for guaranteed loan network risk management | |
Jin et al. | A data-driven approach to predict default risk of loan for online peer-to-peer (P2P) lending | |
Wei et al. | Discovering bank risk factors from financial statements based on a new semi‐supervised text mining algorithm | |
Shi et al. | A novel approach for reducing attributes and its application to small enterprise financing ability evaluation | |
Chye et al. | Credit scoring using data mining techniques | |
Gao et al. | The performance of the P2P finance industry in China | |
Ereiz | Predicting default loans using machine learning (OptiML) | |
CN112417176B (en) | Method, equipment and medium for mining implicit association relation between enterprises based on graph characteristics | |
Liou et al. | Predicting business failure under the existence of fraudulent financial reporting | |
Li et al. | Theory and application of artificial intelligence in financial industry | |
CN115526700A (en) | Risk prediction method and device and electronic equipment | |
Olasehinde‐Williams et al. | Unveiling the causal impact of US economic policy uncertainty on exchange market pressure of African economies | |
CN117010914A (en) | Identification method and device for risk group, computer equipment and storage medium | |
Chen et al. | Financial distress prediction using data mining techniques | |
Huang et al. | Domain adaptation approach for credit risk analysis | |
TWM583089U (en) | Smart credit risk assessment system | |
Cucaro | Measuring the" health" of Italian SMEs with insolvency prediction models Z'-ScoreM and D-Score: Measuring default index of Italian SMEs | |
Mittal et al. | A study on credit risk assessment in banking sector using data mining techniques | |
Muzammil et al. | Determinants for the Adoption of Regulatory Technology (RegTech) Services by the Companies in United Arab Emirates: An MCDM Approach. | |
Abdullah et al. | Mapping crowdfunding research on the web of science database: A bibliometric analysis approach | |
Eko et al. | Improving credit scoring model of mortgage financing with smote methods in sharia banking | |
Li et al. | Textual analysis and detection of financial fraud: Evidence from Chinese manufacturing firms | |
Lee et al. | Application of machine learning in credit risk scorecard | |
Li et al. | Construction of credit evaluation index system for two-stage Bayesian discrimination: an empirical analysis of small Chinese enterprises |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180904 |
|
WD01 | Invention patent application deemed withdrawn after publication |