CN106875271A - Credit assessment method based on walk random on reference man's relational network - Google Patents

Credit assessment method based on walk random on reference man's relational network Download PDF

Info

Publication number
CN106875271A
CN106875271A CN201710087115.2A CN201710087115A CN106875271A CN 106875271 A CN106875271 A CN 106875271A CN 201710087115 A CN201710087115 A CN 201710087115A CN 106875271 A CN106875271 A CN 106875271A
Authority
CN
China
Prior art keywords
reference man
credit
node
credit value
man
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
Application number
CN201710087115.2A
Other languages
Chinese (zh)
Inventor
彭云
徐如志
赵华伟
吴妮
刘晶
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Qilu University of Technology
Original Assignee
Qilu University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Qilu University of Technology filed Critical Qilu University of Technology
Priority to CN201710087115.2A priority Critical patent/CN106875271A/en
Publication of CN106875271A publication Critical patent/CN106875271A/en
Pending legal-status Critical Current

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
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/03—Credit; Loans; Processing thereof

Landscapes

  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

Credit assessment method based on walk random on reference man's relational network of the invention, including：A) sets up reference man's relational network；B) is defined into side neighbours；C) defines side neighbours；D) assigns initial credit value；E) calculates credit value；F) judges whether credit value no longer changes；G) judges whether to meet lending condition.Credit assessment method of the invention, during the credit appraisal to borrower, it is placed in reference man's relational network as reference man, during its credit value is calculated, the credit score do not beaten to it only in accordance with reference man, but also consider the credit situation of the reference man given a mark to it to enter, if the reference man given a mark to it is often given a mark to others and marking is higher, the reference man is probably one and often cheats loan（Credit is very poor）People, then its marking weight will be lower, it is ensured that the accuracy of credit appraisal result.

Description

Credit assessment method based on walk random on reference man's relational network
Technical field
The present invention relates to a kind of credit assessment method, in particular, more particularly to one kind is based on reference man's relational network The credit assessment method of upper walk random.
Background technology
When borrower (such as legal person, natural person) applies providing a loan, it is several that creditor (such as bank) generally needs borrower to submit to Individual reference man (Trade Referee).Behavior of the reference man according to borrower in conventional commercial intercourse is come to borrower's Credit is given a mark.The marking of reference man is higher, and creditor agrees to that the probability of lending (such as business's loan, housing loan) is bigger.
For many years, academia and industrial circle propose many credit assessment methods.Typically have based on expert judgments Credit assessment method, the credit assessment method based on statistical analysis and the credit assessment method based on artificial intelligence.
A) credit assessment method based on expert judgments
The method is the method for comparing early stage.The information that expert manual review borrower is submitted to, it is main to consider 5 dimensions Degree.The refund wish of borrower；The loan repayment capacity of borrower；The net assets of borrower；Guaranty；Also commercioganic environment because Element.But the method efficiency of expert's artificial judgment is too low, it is impossible to process the application of a large amount of borrowers.
B) credit assessment method based on statistical analysis
With the fast development of the fast development of personal finance, especially credit card, academia and industry are proposed and are based on The credit assessment method of statistical analysis.The FICO in the U.S. is the most commonly used method of current application.The method investigates borrower ten The information of several dimensions.Such as refund history, credit accounts number, credit history, the type of credit accounts.These aspects Information statistical model is built as the characteristic vector of credit, the letter of borrower is evaluated by statistical model With.
C) credit assessment method based on artificial intelligence
Recently as the fast development of internet, artificial intelligence and financial technology (Fintech), academia and industry are carried The credit assessment method based on artificial intelligence is gone out.Its thought is magnanimity information that borrower is left on the internet as spy Levy, train a model for artificial intelligence.The method can contemplate borrower's educational background, current consumption history, water power and warm up and pay on time Expense history, rent are paid the fees the thousands of dimensions such as history, net purchase behavior on time.
But existing credit assessment method has a drawback, reference man is according to borrower in conventional commercial intercourse Behavior given a mark come the credit to borrower, but reference man the scope that does not account for of credit.Reference man is probably One people for often cheating loan, his credit is very poor, and his marking should be insincere.Existing method does not consider such case, The result of credit appraisal is caused to be forbidden.
The content of the invention
Shortcoming in order to overcome above-mentioned technical problem of the invention, there is provided one kind is based on being overflow at random on reference man's relational network The credit assessment method of step.
Credit assessment method based on walk random on reference man's relational network of the invention, it is particular in that, according to It is secondary to be realized by following steps：
A) sets up reference man's relational network, and reference man's relational network, wherein reference man v are represented with figure G=(V, E)iIt is V In a node, V={ v1,v2,...,vi,...,vn1, n1 is the number of reference man；Marking relation between reference man is Side, if reference man v gives a mark to reference man u, the marking relation between them is represented with directed edge e (v, u), and e (v, u) is in E A directed edge, the fraction that v is beaten to u is the weight on side e (v, u), is represented with s (v, u)；Wherein, v ∈ V, u ∈ V, and v It is not same reference man with u；
B) is defined into side neighbours, fixed in figure G=(V, E) if reference man v has the marking behavior to reference man u Adopted node v is that node u enters side neighbours, node u it is all enter side neighbours with set NinU () represents；V ∈ V, u ∈ V, and v and u It is not same reference man；
C) defines side neighbours, correspondingly, if reference man v has the marking behavior to reference man u, in figure G= Node u is that node v goes out side neighbours defined in (V, E), node v it is all go out side neighbours with set NoutV () represents；v∈V、u ∈ V, and v and u is not same reference man；
D) assigns initial credit value, to the corresponding node v of reference man in figure G=(V, E)iAssign one it is equal initial Credit value trank (vi), i=1,2 ..., n1；
E) calculates credit value, for each node in figure G=(V, E), recalculates its credit using formula (1) Value：
Wherein, node v enters side neighbours for node u, and v ' goes out side neighbours for v's；
Perform step f)；
F) judges whether credit value no longer changes, and whether the node credit value after judgement is recalculated is sent out compared with before Change has been given birth to, if there is the node of credit value changes, then step e) has been performed；If the credit value of all nodes no longer becomes Change, then perform step g)；
G) judges whether to meet lending condition, the final credit value trank (v of the reference man that will be investigatedi) and creditor The lending credit threshold of setting is compared, if the credit value of reference man is more than lending credit threshold, provides satisfaction and makes loans The suggestion of condition；If the credit value of reference man is less than lending credit threshold, the suggestion for being unsatisfactory for lending condition is given.
Credit assessment method based on walk random on reference man's relational network of the invention, in step d), assigns reference The initial credit value of people is that the credit value of calculating in 100, step e) is accurate to 2 significant digits.
Credit assessment method based on walk random on reference man's relational network of the invention, if assigning the initial of reference man Credit value is X, then the lending credit threshold in step g) is (1+ α) X, α ＞ 0.
The beneficial effects of the invention are as follows：Credit assessment method of the invention, during the credit appraisal to borrower, It is placed in reference man's relational network as reference man, during its credit value is calculated, does not give it only in accordance with reference man The credit score beaten, but also the credit situation of the reference man given a mark to it is considered to enter, if the reference given a mark to it People often gives a mark to others and gives a mark higher, and the reference man is probably a people for often cheating loan (credit is very poor), then it is given a mark Weight will be lower, it is ensured that the accuracy of credit appraisal result.
Brief description of the drawings
Reference man's relational network figure that Fig. 1 is set up by embodiments of the invention；
Fig. 2 is the schematic diagram of each node tax initial credit value in embodiments of the invention；
Fig. 3 is the result schematic diagram after embodiments of the invention first time iteration；
Fig. 4 is the result schematic diagram after second iteration of embodiments of the invention；
Fig. 5 is the result schematic diagram after embodiments of the invention third time iteration；
Fig. 6 is the result schematic diagram after the 4th iteration of embodiments of the invention；
Fig. 7 is the result schematic diagram after the 5th iteration of embodiments of the invention.
Specific embodiment
The invention will be further described with embodiment below in conjunction with the accompanying drawings.
As shown in figure 1, giving reference man's relational network figure that embodiments of the invention are set up, had in the embodiment 4 reference man v0、v1、v2、v3, its interior joint v0Give node v1、v2、v310 scores, 10 points, 0.1 point are made respectively；v1Only give v0、v3Beat Point, grade respectively 10 points, 0.1 point, v3Only give v1、v2Marking, grade respectively 10 points, 0.1 point, v2Only give v0Beat Point, it is 10 points to grade.
Understand：v0Enter side neighbours Nin(v0) it is v1And v2, go out side neighbours Nout(v0) it is v1、v2、v3；v1Enter side neighbours Nin(v1) it is v0And v3, go out side neighbours Nout(v1) also it is v0And v3；v2Enter side neighbours Nin(v2) it is v0And v3, go out side neighbours Nout (v2) it is v0；v3Enter side neighbours Nin(v3) it is v0And v1, go out side neighbours Nout(v3) it is v1And v2。
As shown in Fig. 2 each node assigns the schematic diagram of initial credit value in giving embodiments of the invention, it is seen then that node v0、v1、v2、v3Initial credit value trank (vi) it is 100, i=1,2,3,4.Then using following formula (1) to node v0、v1、v2、v3Credit value recalculated：
Wherein, node v enters side neighbours for node u, and v ' goes out side neighbours for v's.
In first time iterative process, to node v0Credit value when being calculated, v0Enter side neighbours Nin(v0) it is v1With v2, v1、v2To v0Marking be 10 points.v1Go out side neighbours Nout(v1) it is v0And v3, v1To v0、v3Marking be respectively 10 points, 0.1 point；v2Go out side neighbours Nout(v2) it is v0, v2To v0Marking be 10 points, therefore bringing formula (1) into can calculate egress v0's Credit value：
Similarly, first time iteration posterior nodal point v can be calculated1、v2And v3Credit value be respectively 148.76,50.74 and 1.48。
As shown in figure 3, giving the result schematic diagram after embodiments of the invention first time iteration, first time iteration Afterwards, v0、v1、v2、v3Credit value 199.00,148.76,50.74,1.48, there occurs change relative to former credit value 100, Therefore second iterative calculation should be carried out.
As shown in figure 4, give the result schematic diagram after second iteration of embodiments of the invention, second iteration it Afterwards, v0、v1、v2、v3Credit value be respectively 198.03,100.48,99.02,2.46, its relative to former credit value 199.00, 148.76th, 50.74,1.48 change is there occurs, therefore third time iterative calculation should be carried out.
As shown in figure 5, give the result schematic diagram after embodiments of the invention third time iteration, third time iteration it Afterwards, v0、v1、v2、v3Credit value be respectively 198.51,100.96,98.55,1.98, its relative to former credit value 198.03, 100.48th, 99.02,2.46 change is there occurs, therefore the 4th iterative calculation should be carried out.
As shown in fig. 6, give the result schematic diagram after the 4th iteration of embodiments of the invention, the 4th iteration it Afterwards, v0、v1、v2、v3Credit value be respectively 198.51,100.72,98.78,1.48, v1、v2、v3Credit value 100.72, 98.78th, 1.48 change is there occurs relative to former credit value 100.96,98.55,1.98, therefore the 5th iterative calculation should be carried out.
As shown in fig. 7, give the result schematic diagram after the 5th iteration of embodiments of the invention, the 5th iteration it Afterwards, v0、v1、v2、v3Credit value be respectively 198.51,100.72,98.78,1.48, its credit value does not change, therefore Terminate iterative calculation.
If lending credit threshold is 100, node v0、v1Corresponding reference man meets lending condition.If lending credit Threshold value is 150, then only node v0Corresponding reference man meets lending condition.

Claims (3)

1. a kind of credit assessment method based on walk random on reference man's relational network, it is characterised in that pass sequentially through following Step is realized：
A) sets up reference man's relational network, and reference man's relational network, wherein reference man v are represented with figure G=(V, E)iIt is in V Individual node, V={ v1,v2,...,vi,...,vn1, n1 is the number of reference man；Marking relation between reference man is side, if ginseng Examine people v to give a mark reference man u, then represent the marking relation between them with directed edge e (v, u), e (v, u) is one in E to be had Xiang Bian, v are the weight on side e (v, u) to the fraction that u beats, and are represented with s (v, u)；Wherein, v ∈ V, u ∈ V, and v and u is not Same reference man；
B) is defined into side neighbours, if reference man v has the marking behavior to reference man u, is saved defined in figure G=(V, E) Point v is that node u enters side neighbours, node u it is all enter side neighbours with set NinU () represents；V ∈ V, u ∈ V, and v and u is not Same reference man；
C) defines side neighbours, correspondingly, if reference man v has the marking behavior to reference man u, in figure G=(V, E) Defined in node u be that node v goes out side neighbours, node v it is all go out side neighbours with set NoutV () represents；V ∈ V, u ∈ V, and V and u is not same reference man；
D) assigns initial credit value, to the corresponding node v of reference man in figure G=(V, E)iAssign an equal initial credit value trank(vi), i=1,2 ..., n1；
E) calculates credit value, for each node in figure G=(V, E), recalculates its credit value using formula (1)：
t r a n k ( u ) = Σ v ∈ N i n ( u ) ( t r a n k ( v ) · s ( v , u ) Σ v ′ ∈ N o u t ( v ) s ( v , v ′ ) ) - - - ( 1 )
Wherein, node v enters side neighbours for node u, and v ' goes out side neighbours for v's；
Perform step f)；
F) judges whether credit value no longer changes, and whether the node credit value after judgement is recalculated there occurs compared with before Change, if there is the node of credit value changes, then performs step e)；If the credit value of all nodes no longer changes, Perform step g)；
G) judges whether to meet lending condition, the final credit value trank (v of the reference man that will be investigatedi) set with creditor Lending credit threshold be compared, if the credit value of reference man be more than lending credit threshold, be given and meet lending condition Suggestion；If the credit value of reference man is less than lending credit threshold, the suggestion for being unsatisfactory for lending condition is given.
2. the credit assessment method based on walk random on reference man's relational network according to claim 1, its feature exists In：In step d), the initial credit value for assigning reference man is that the credit value of calculating in 100, step e) is accurate to two after decimal point Position.
3. the credit assessment method based on walk random on reference man's relational network according to claim 1 and 2, its feature It is：If the initial credit value for assigning reference man is X, then the lending credit threshold in step g) is (1+ α) X, α ＞ 0.
CN201710087115.2A 2017-02-17 2017-02-17 Credit assessment method based on walk random on reference man's relational network Pending CN106875271A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710087115.2A CN106875271A (en) 2017-02-17 2017-02-17 Credit assessment method based on walk random on reference man's relational network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710087115.2A CN106875271A (en) 2017-02-17 2017-02-17 Credit assessment method based on walk random on reference man's relational network

Publications (1)

Publication Number Publication Date
CN106875271A true CN106875271A (en) 2017-06-20

Family

ID=59167143

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710087115.2A Pending CN106875271A (en) 2017-02-17 2017-02-17 Credit assessment method based on walk random on reference man's relational network

Country Status (1)

Country Link
CN (1) CN106875271A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107871278A (en) * 2017-08-04 2018-04-03 平安普惠企业管理有限公司 The method and storage medium of server, customer relationship network display
CN107871277A (en) * 2017-07-25 2018-04-03 平安普惠企业管理有限公司 The method and computer-readable recording medium that server, customer relationship are excavated
CN108256990A (en) * 2017-07-25 2018-07-06 平安普惠企业管理有限公司 Server, the method for indicating risk and computer readable storage medium
WO2020062641A1 (en) * 2018-09-26 2020-04-02 深圳壹账通智能科技有限公司 Method for identifying user role, and user equipment, storage medium, and apparatus for identifying user role

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107871277A (en) * 2017-07-25 2018-04-03 平安普惠企业管理有限公司 The method and computer-readable recording medium that server, customer relationship are excavated
CN108256990A (en) * 2017-07-25 2018-07-06 平安普惠企业管理有限公司 Server, the method for indicating risk and computer readable storage medium
CN108256990B (en) * 2017-07-25 2021-07-23 平安普惠企业管理有限公司 Server, risk prompting method and computer readable storage medium
CN107871278A (en) * 2017-08-04 2018-04-03 平安普惠企业管理有限公司 The method and storage medium of server, customer relationship network display
CN107871278B (en) * 2017-08-04 2021-08-31 平安普惠企业管理有限公司 Server, client relationship network display method and storage medium
WO2020062641A1 (en) * 2018-09-26 2020-04-02 深圳壹账通智能科技有限公司 Method for identifying user role, and user equipment, storage medium, and apparatus for identifying user role

Similar Documents

Publication Publication Date Title
CN106875271A (en) Credit assessment method based on walk random on reference man's relational network
Rosser From catastrophe to chaos: a general theory of economic discontinuities
CN107194723B (en) Bidirectional matching recommendation method for borrowing items and lenders in network micropayment
Boughton The IMF and the force of history: Ten events and ten ideas that have shaped the institution
Thurow Fortune favors the bold: What we must do to build a new and lasting global prosperity
CN107993140A (en) A kind of personal credit's methods of risk assessment and system
US6951008B2 (en) Evidential reasoning system and method
Looney The Jamaican economy in the 1980s: economic decline and structural adjustment
WO2003094034A1 (en) Method of training a neural network and a neural network trained according to the method
Casey et al. A Simple Theory of Complex Valuation
Voth With a bang, not a whimper: Pricking Germany's “stock market bubble” in 1927 and the slide into depression
Jagielska et al. Neural network for predicting the performance of credit card accounts
CN112200656A (en) On-line pre-approval method, device, medium and electronic equipment for house loan
Wray A Great Leap Forward: Heterodox economic policy for the 21st century
US20020198819A1 (en) Method and apparatus for risk based pricing
Lewis The origins of the sub-prime crisis: Inappropriate policies, regulations, or both?
Dymski et al. Race, power, and the subprime/foreclosure crisis: a mesoanalysis
CN111507825A (en) Bidirectional recommendation method based on game multi-objective optimization on P2P platform
Krantz et al. A mathematical odyssey
Tirado Complex network for a crisis contagion on an interbank system
Mellyn Broken markets: a user's guide to the post-finance economy
CN114004693A (en) Method for determining residual value evaluation of poor asset of consumption credit
Diehr The Looming Threat Posed to Student Loan Lenders and Servicers by State-Level Actors in an Era of Federal Regulatory Remission
Gallagher et al. Reforming Bretton Woods institutions to achieve climate change and development goals
Setiawan et al. Investment Decisions and Investor Experience in Choosing Financial Technology (Fintech) Services: Peer-To-Peer Lending

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: 20170620

WD01 Invention patent application deemed withdrawn after publication