CN107679982A - A kind of credit card risk checking method based on point process - Google Patents
A kind of credit card risk checking method based on point process Download PDFInfo
- Publication number
- CN107679982A CN107679982A CN201710904426.3A CN201710904426A CN107679982A CN 107679982 A CN107679982 A CN 107679982A CN 201710904426 A CN201710904426 A CN 201710904426A CN 107679982 A CN107679982 A CN 107679982A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- mtd
- consumption
- credit card
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
-
- 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
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Accounting & Taxation (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Educational Administration (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Technology Law (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of credit card risk checking method based on point process, collect first and arrange user credit card consumption data, then user credit card usage behavior is established by point process and changes over time model, user's dynamic consumption habit is integrated into the model of point process, user's dynamic consumption matrix is established as user credit card consumption habit model, with the user credit card consumption habit model prediction user following credit card trade time to be occurred, risk factor is exported finally by risk forecast model, model noise resisting ability is stronger, also human behavior rule is more conformed to, the degree of accuracy is higher.
Description
Technical field
The invention belongs to credit card risk supervision technical field, more specifically, is related to a kind of letter based on point process
With card risk checking method,
Background technology
With the rapid development of economy, increasing people is consumed by credit card.Credit card is widely used as
While user's offer consumption facilitates, larger credit card application risk is also brought.
Credit card application risk mainly includes two aspects：1) credit card fraud, robber's brush, the loss of personal credit card, emit
Neck, steal brush, personation apply, forged credit card etc., as it is non-in person the reason for caused by personal credit card risk.Personal letter in recent years
Belong to case occurred frequently with card fraud case, be often only such case and cause damage altogether up to more than one hundred million members, credit card fraud into
The most important thing paid close attention to for bank and supervision department.2) personal to be gambled overseas, money laundering etc. passes through illegal caused by credit card
Banking operation, fall within credit card risk.With the lifting of economic level, the individual wealth of some people is lifted rapidly, due to
Gambling belongs to illegal act at home, and some criminals utilization supervises leak and utilizes the illegal arbitrage of credit card abroad, forges
Transaction etc., malversation carries out money-laundering overseas by credit card.These overseas illegally use the Behavioral availability of credit card
Supervision missing, huge economic loss is brought to country.Strengthen credit card supervision, take precautions against credit card risk, establish reasonable, strong
The credit card employment mechanism of health, the illegal acts overseas such as personal consumption, prevention money laundering, gambling are may advantageously facilitate, establishing specification has
The credit card use environment of sequence.
Credit card risk supervision has become an important hot issue, has attracted the association areas such as economy, finance
The research of experts and scholars, in recent years due to the development of big data technology, the researcher of computer realm is also to credit card risk
Detection conducts in-depth research.
Traditional credit card risk monitoring method is when preventing the illegal act such as gambling, money laundering overseas, based on spending amount
Single factors establish certain rule, when the consumption abroad put into effect such as recent supervision department is more than 1000 yuans, it is necessary on
Report.Credit card risk monitoring method based on spending amount, do not account for everyone level of consumption, consumption habit, consumption with
The factors such as the change of time, the method for single solution for diverse problems is taken to be supervised, effect is poor, in terms of supervision and credit card using flexible not
Balance well can be realized.In terms of personal credit card Risk Monitoring, the mechanism of the SMS notification after generally use consumption, not
Giving warning in advance property can be carried out to consumer behavior to inform, after credit card fraud is had occurred that, user just has found that credit card is taken advantage of
The hysteresis of swindleness information causes economic loss.
The content of the invention
It is an object of the invention to overcome the deficiencies in the prior art, proposes a kind of credit card risk supervision based on point process
Method, by establishing the model being accustomed to based on customer consumption, effectively analyze current credit card consumer behavior and be accustomed to customer consumption
Between difference, to improve the pre-alerting ability of credit card risk, reduce credit card fraud risk.
For achieving the above object, the credit card risk checking method of the invention based on point process, it is characterised in that bag
Include following steps：
(1), mapping structure credit card purchase tuple
The credit card purchase data of user are collected, and user credit card consumption data arrange clearly according to analysis demand
Wash, the consumption information mapping structure credit card purchase tuple (position class of consumer behavior will be reflected in user credit card consumption data
Not, section, consumption time are consumed), and it is designated as (lk,ck,tk), wherein, lkFor the position classification of kth time consumption, ckDisappear for kth time
The consumption section taken, tkFor the consumption time of kth time consumption；
(2) the user credit card consumption habit model based on point process, is built
Establish user's dynamic consumption matrix A (t) and be used as user credit card consumption habit model：
Wherein, λij(t) represent current time t user to i-th of position classification, j-th consumption section preference, i=1,
2,…,NL, j=1,2 ..., NC, NL、NCRespectively position classification quantity, consumption section quantity, λij(t) it is：
Wherein, λij(t0) represent t0Before moment, user is in i-th of position classification, the consumption time in j-th of consumption section
Number NijAccount for the ratio of all consumption number of times, λij(t0) be：
Wherein, NmnRepresent t0Before moment, user is in m-th of position classification, n-th of consumption number of times for consuming section；
K represents the number consumed to current time t user, Ik(Li,Cj) represent that kth time is consumed in i-th of position class
Not, the indicator function in section is consumed at j-th, its value is：
Wherein, LiRepresent i-th of position classification, CjRepresent j-th of consumption section；
The time effects of kth time consumption are represented, are a time attenuation functions, τ is decay factor；
(3), to the time interval T ' for the credit card purchase behavior that will occurijIt is predicted
Time Density function f of the user to preferenceij(t) it is：
The time interval T ' then consumed next time in i-th of position classification, j-th of consumption sectionijFor：
Wherein, before t' is current time t, the last time carries out credit card in i-th of position classification, j-th of consumption section
The time of consumption；
(4) risk assessment, is carried out to the credit card purchase behavior that will currently occur
When the current credit card purchase information that will occur is transmitted back into credit card centre, current consumption position is calculated first
The gap with spending amount and consumption preferences is put, the Part I being expressed as in formula (7), credit card is then calculated and currently will
I-th of the position classification occurred, last the disappearing with consumption section with position classification of exchange hour distance for consuming section j-th
Time taking time interval is actual time interval Tij, then calculate reflection actual time interval TijWith predicted time interval T 'ij
Between gap be expressed as Part II in formula (7), computing metric D (t)：
Wherein, the probability that i-th of position classification will occur for current credit card, the exchange in j-th of consumption section occurs
Represented with 1, λ 'ij(t) section preference normalized value is consumed to i-th of position classification, j-th for user：
Wherein, u is the average of all elements value in user's dynamic consumption matrix A (t), and σ is user's dynamic consumption matrix A
(t) standard deviation of all elements value in；
Finally export risk factor Risk：
If risk factor exceeds threshold value, send and confirm that short message confirms consumer behavior to user, and report supervision department to put on record.
The object of the present invention is achieved like this.
Credit card risk checking method of the invention based on point process, collect first and arrange user credit card consumption number
According to, user credit card usage behavior is then established by point process and changes over time model, by user's dynamic consumption habit merge
Enter in the model of point process, user's dynamic consumption matrix is established as user credit card consumption habit model, with user credit
The card consumption habit model prediction user credit card trade time to be occurred in future, wind is exported finally by risk forecast model
Dangerous coefficient, directly as the result of risk alarm.The present invention technical thought compare with conventional credit card risk checking method, point
Analysed user credit card using preference and spending amount, consumption position classification, consumption time between relation, and establish base
Preference of dynamic model is used in the credit card of point process, the probability and the time of generation that prediction same type credit card trade occurs,
Risk evaluation model is finally established, assesses the risk of credit card trade, model noise resisting ability is stronger, also more conforms to mankind's row
It is higher for rule, the degree of accuracy.
Brief description of the drawings
Fig. 1 is credit card risk checking method a kind of embodiment flow chart of the invention based on point process；
Fig. 2 is the credit card purchase one embodiment schematic diagram of tuple of mapping structure shown in Fig. 1；
Fig. 3 is the timing diagram of the consumption habit model of user credit card shown in Fig. 1；
Fig. 4 is that the credit card purchase behavior that will currently occur shown in Fig. 1 carries out the model framework chart of risk assessment；
Fig. 5 is the embodiment flow chart of the credit card risk checking method of the invention based on point process.
Embodiment
The embodiment of the present invention is described below in conjunction with the accompanying drawings, so as to those skilled in the art preferably
Understand the present invention.Requiring particular attention is that in the following description, when known function and the detailed description of design perhaps
When can desalinate the main contents of the present invention, these descriptions will be ignored herein.
Fig. 1 is credit card risk checking method a kind of embodiment flow chart of the invention based on point process.
In the present embodiment, as shown in figure 1, the credit card risk checking method of the invention based on point process includes following step
Suddenly：
Step S1：Mapping structure credit card purchase tuple
User credit card consumption data includes consumption position when user uses credit card, spending amount, consumption time etc.
Information, user are only relevant with consuming the consumption information such as position, spending amount, consumption time using the consumer behavior of credit card.Choosing
Consumption information each time is selected to map to form a credit card purchase tuple cardk=(lk,ck,tk) it is a Card tuple, its
In, lkFor the position classification of kth time consumption, ckFor the consumption section of kth time consumption, tkFor the consumption time of kth time consumption.Position
The division for putting classification determines according to expertise, uses NLThe granularity of division for representing position classification is position categorical measure, position class
Shop, hospital, market, luxury goods shop, the domestic, N such as overseas can be divided intoLClass.Consuming section can be according to the volume of user credit card
The degree amount of money is divided, and the granularity of division according to supervision demand and is actually needed determination, uses NCRepresent that the granularity of division is consumed
Section quantity.Consumption time can be in units of day.
Mapping structure credit card purchase tuple schematic diagram is as shown in Figure 2.Built by this mapping, can will be discrete
Credit card purchase data are expressed as regular credit card tuple Card, and each credit card purchase is carried out into mapping structure, worked as
Credit card is designated as CT using track Card Trajectory before preceding time t, represents as shown in (10)：
CT={ Card1,Card2,…,Cardk,…CardK} (10)
Step S2：Build the user credit card consumption habit model based on point process
Point process can be portrayed in the unit interval, the influence expression formula of event frequency, can portray the influence of event
Factor, dynamic catch the Behavior preference of user, can preferably portray user in using credit card, position classification, consumption area
Between and consumption time relation.
After have passed through credit card purchase data preparation cleaning and mapping structure, the credit card for obtaining each user uses rail
After mark, establish user's dynamic consumption matrix A (t) and be used as user credit card consumption habit model, consumption matrix A (t) represents user
It is N to each position classification, the preference in each consumption sectionL×NCThe matrix of size, and matrix element λij(t) it is user
The preference of each position classification, each consumption section is dynamically updated over time, user's dynamic consumption matrix A (t) is：
Wherein, λij(t) represent current time t user to i-th of position classification, j-th consumption section preference, i=1,
2,…,NL, j=1,2 ..., NC, NL、NCRespectively position classification quantity, consumption section quantity, λij(t) it is：
Wherein, λij(t0) represent t0Before moment, user is in i-th of position classification, the consumption time in j-th of consumption section
Number NijAccount for the ratio of all consumption number of times, λij(t0) be：
Wherein, NmnRepresent t0Before moment, user is in m-th of position classification, n-th of consumption number of times for consuming section.
All position classifications, the λ in all consumption sectionsij(t0) composition customer consumption matrix initial matrix A (0).
t0The credit card purchase occurred after moment, dynamic update customer consumption matrix initial matrix A (0), more new formula
For：
K represents the number consumed to current time t user, and it is discrete counts process that point process, which is based on Poisson process, meter
Number function uses indicator function Ik(Li,Cj) withProduct,The time effects of kth time consumption are represented, when being one
Between attenuation function, τ is decay factor, Ik(Li,Cj) represent that kth time is consumed in i-th of position classification, consume section at j-th
Indicator function, its value are：
Wherein, LiRepresent i-th of position classification, CjRepresent j-th of consumption section.
Based on user credit card consumption habit model can dynamically update customer consumption matrix, catch user and use credit
To consumption position, the preference of spending amount during card consumption.Meanwhile by two above structure based on user credit card consume
It is accustomed to model, can be good at portraying the past consumption preferences of user and dynamic catches the consumption preferences of user.The dynamic process
The influence that time attenuation function portrays same type consumption is with the addition of, feature that user credit card consumer behavior changes over time is dynamic
State property is as shown in Figure 3.
Step S3：To the time interval T ' for the credit card purchase behavior that will occurijIt is predicted
, it is necessary to dynamic prediction user next time i-th after the user credit card consumption habit model based on point process is obtained
Individual position classification, j-th of consumption section carry out the time interval T ' of credit card purchase behaviorij.User is estimated by formula (2)
To the Time Density function f of preferenceij(t)：
The time interval T ' then consumed next time in i-th of position classification, j-th of consumption sectionijFor：
Wherein, before t' is current time t, the last time carries out credit card in i-th of position classification, j-th of consumption section
The time of consumption.
Step S4：Risk assessment is carried out to the credit card purchase behavior that will currently occur
User credit card consumption habit model of the present invention structure based on point process, which can also be analyzed, to be drawn and currently will
The credit card whether the credit card purchase behavior of generation meets user uses habit.The inspection of credit card risk is also constructed in this step
Survey model, the normal consumer behavior that can be user with the imminent credit card purchase behavior of effective detection (judgement) user,
Or the consumer sale behavior with risk.
Dynamic Maintenance renewal user's dynamic consumption matrix A (t), and predicted by formula (5) next time in i-th of position class
Not, the time interval T ' that j-th of consumption section is consumedij, the current credit card transaction information that will occur is transmitted back to
During credit card centre, calculate first current imminent i-th of the position classification of credit card, j-th consumption section consumption when
Between the distance last time with position classification be actual time interval T with the time interval of consumption time in consumption sectionij, then calculate
This two distance metric values of the gap of current consumption and user preference, the real world of customer consumption and the gap of predicted time
With D (t)：
Wherein, the probability that the consumption of i-th of position classification, j-th of consumption section is occurred will occur for current credit card
Represented with 1, λ 'ij(t) section preference normalized value is consumed to i-th of position classification, j-th for user：
Wherein, u is the average of all elements value in user's dynamic consumption matrix A (t), and σ is user's dynamic consumption matrix A
(t) standard deviation of all elements value in；
Finally export risk factor Risk：
If risk factor exceeds threshold value, send and confirm that short message confirms consumer behavior to user, and report supervision department to put on record.
Risk factor Risk is the percentage between 0% to 100%, directly represents the risk system of current consumption behavior
Number.After risk supervision, new credit card purchase data are continuously added in user's dynamic consumption matrix A (t).
Credit card risk checking method of the invention based on point process can be good at detection credit card and usurp, steal brush, emit
Overseas disappeared with, money laundering, the big amount of money of high frequency time using credit card application risk, the user credit cards based on point process such as credits card
Expense custom model can analyze the consumption preferences that user uses credit card, and the different of credit card custom is used so as to find to run counter to user
Normal credit card purchase behavior.When suspicious credit card purchase behavior will occur for user, if what risk assessment detected
Risk factor higher i.e. risk factor is sent and confirms that short message gives user's confirmation consumer behavior, and report supervision department when exceed threshold value
Put on record.Risk assessment processes are as shown in Figure 4.
The embodiment flow of credit card risk checking method of the invention described above based on point process is as shown in Figure 5.This hair
Bright position classification, spending amount and the consumption time extracted first in user credit card consumption data forms credit card track, so
User is established on position classification, spending amount by building user credit card usage time model based on point process afterwards
Preference relation, user's dynamic consumption matrix A (t) of updating maintenance user, draw user for different consumption classifications (including positions
Classification, consumption section) prediction, and the consumption time of same type next time is predicted, finally by measurement actual value and predicted value
Otherness, build one's credit card risk evaluation model, the risk factor of credit card trade behavior that output will occur.If
Beyond threshold value, then issue the user with confirmation and put on record to supervision department, after obtaining user's confirmation, perform credit card trade.
Although the illustrative embodiment of the present invention is described above, in order to the technology of the art
Personnel understand the present invention, it should be apparent that the invention is not restricted to the scope of embodiment, to the common skill of the art
For art personnel, if various change in the spirit and scope of the present invention that appended claim limits and determines, these
Change is it will be apparent that all utilize the innovation and creation of present inventive concept in the row of protection.
Claims (1)
1. a kind of credit card risk checking method based on point process, it is characterised in that comprise the following steps：
(1), mapping structure credit card purchase tuple
The credit card purchase data of user are collected, and arrangement cleaning is carried out to user credit card consumption data according to analysis demand,
By reflect in user credit card consumption data consumer behavior consumption information mapping structure credit card purchase tuple (position classification,
Consume section, consumption time), and it is designated as (lk,ck,tk), wherein, lkFor the position classification of kth time consumption, ckFor kth time consumption
Consumption section, tkFor the consumption time of kth time consumption；
(2) the user credit card consumption habit model based on point process, is built
Establish user's dynamic consumption matrix A (t) and be used as user credit card consumption habit model：
<mrow>
<mi>A</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "(" close = ")">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>&lambda;</mi>
<mn>11</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
<mtd>
<mrow>
<msub>
<mi>&lambda;</mi>
<mn>12</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
<mtd>
<mn>...</mn>
</mtd>
<mtd>
<mrow>
<msub>
<mi>&lambda;</mi>
<mrow>
<mn>1</mn>
<msub>
<mi>N</mi>
<mi>C</mi>
</msub>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>&lambda;</mi>
<mn>21</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
<mtd>
<mrow>
<msub>
<mi>&lambda;</mi>
<mn>22</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
<mtd>
<mn>...</mn>
</mtd>
<mtd>
<mrow>
<msub>
<mi>&lambda;</mi>
<mrow>
<mn>2</mn>
<msub>
<mi>N</mi>
<mi>C</mi>
</msub>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
<mtd>
<mo>.</mo>
</mtd>
<mtd>
<mrow></mrow>
</mtd>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
<mtd>
<mo>.</mo>
</mtd>
<mtd>
<mn>...</mn>
</mtd>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo>.</mo>
</mtd>
<mtd>
<mo>.</mo>
</mtd>
<mtd>
<mrow></mrow>
</mtd>
<mtd>
<mo>.</mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>&lambda;</mi>
<mrow>
<msub>
<mi>N</mi>
<mi>L</mi>
</msub>
<mn>1</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
<mtd>
<mrow>
<msub>
<mi>&lambda;</mi>
<mrow>
<msub>
<mi>N</mi>
<mi>L</mi>
</msub>
<mn>2</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
<mtd>
<mn>...</mn>
</mtd>
<mtd>
<mrow>
<msub>
<mi>&lambda;</mi>
<mrow>
<msub>
<mi>N</mi>
<mi>L</mi>
</msub>
<msub>
<mi>N</mi>
<mi>C</mi>
</msub>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, λij(t) represent current time t user to i-th of position classification, j-th consumption section preference, i=1,2 ...,
NL, j=1,2 ..., NC, NL、NCRespectively position classification quantity, consumption section quantity, λij(t) it is：
<mrow>
<msub>
<mi>&lambda;</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mi>&lambda;</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>t</mi>
<mn>0</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>+</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>k</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>K</mi>
</munderover>
<msub>
<mi>I</mi>
<mi>k</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>L</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>C</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>&times;</mo>
<msup>
<mi>e</mi>
<mrow>
<mi>&tau;</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>-</mo>
<msub>
<mi>t</mi>
<mi>k</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, λij(t0) represent t0Before moment, user is in i-th of position classification, the consumption number of times N in j-th of consumption sectionij
Account for the ratio of all consumption number of times, λij(t0) be：
<mrow>
<msub>
<mi>&lambda;</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>t</mi>
<mn>0</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<msub>
<mi>N</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<msub>
<mi>N</mi>
<mi>L</mi>
</msub>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>n</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<msub>
<mi>N</mi>
<mi>C</mi>
</msub>
</munderover>
<msub>
<mi>N</mi>
<mrow>
<mi>m</mi>
<mi>n</mi>
</mrow>
</msub>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, NmnRepresent t0Before moment, user is in m-th of position classification, n-th of consumption number of times for consuming section；
K represents the number consumed to current time t user, Ik(Li,Cj) represent that kth time is consumed in i-th of position classification,
The indicator function in j-th of consumption section, its value are：
Wherein, LiRepresent i-th of position classification, CjRepresent j-th of consumption section；
The time effects of kth time consumption are represented, are a time attenuation functions, τ is decay factor；
(3), to the time interval T for the credit card purchase behavior that will occurij' be predicted
Time Density function f of the user to preferenceij(t) it is：
<mrow>
<msub>
<mi>f</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mi>&lambda;</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>&times;</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<msubsup>
<mo>&Integral;</mo>
<msup>
<mi>t</mi>
<mo>&prime;</mo>
</msup>
<mi>t</mi>
</msubsup>
<msub>
<mi>&lambda;</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>(</mo>
<mi>s</mi>
<mo>)</mo>
<mi>d</mi>
<mi>s</mi>
<mo>)</mo>
</mrow>
</msup>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
The time interval T then consumed next time in i-th of position classification, j-th of consumption sectionij' be：
<mrow>
<msubsup>
<mi>T</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mo>&prime;</mo>
</msubsup>
<mo>=</mo>
<msubsup>
<mo>&Integral;</mo>
<msup>
<mi>t</mi>
<mo>&prime;</mo>
</msup>
<mi>&infin;</mi>
</msubsup>
<mi>t</mi>
<mo>&times;</mo>
<msub>
<mi>f</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mi>d</mi>
<mi>t</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>6</mn>
<mo>)</mo>
</mrow>
</mrow>
(4) risk assessment, is carried out to the credit card purchase behavior that will currently occur
Will the current credit card purchase information that will occur when being transmitted back to credit card centre, calculate first current consumption position and
The gap of spending amount and consumption preferences, the Part I being expressed as in formula (7), then calculating credit card will currently occur
I-th of position classification, the exchange hour distance in j-th of consumption section it is last with position classification with consumption section consumption when
Between time interval be actual time interval Tij, then calculate reflection actual time interval TijWith predicted time interval Tij' between
Gap is expressed as the Part II in formula (7), computing metric D (t)：
<mrow>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mo>|</mo>
<mfrac>
<mrow>
<mn>1</mn>
<mo>-</mo>
<msubsup>
<mi>&lambda;</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mo>&prime;</mo>
</msubsup>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msubsup>
<mi>&lambda;</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mo>&prime;</mo>
</msubsup>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>|</mo>
<mo>+</mo>
<mo>|</mo>
<mfrac>
<mrow>
<msub>
<mi>T</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>-</mo>
<msubsup>
<mi>T</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mo>&prime;</mo>
</msubsup>
</mrow>
<msubsup>
<mi>T</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mo>&prime;</mo>
</msubsup>
</mfrac>
<mo>|</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, 1 table of the probability that i-th of position classification will occur for current credit card, the exchange in j-th of consumption section occurs
Show, λij' (t) is user to i-th of position classification, j-th of consumption section preference normalized value：
<mrow>
<msubsup>
<mi>&lambda;</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
<mo>&prime;</mo>
</msubsup>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>&lambda;</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>u</mi>
</mrow>
<mi>&sigma;</mi>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, u is the average of all elements value in user's dynamic consumption matrix A (t), and σ is in user's dynamic consumption matrix A (t)
The standard deviation of all elements value；
Finally export risk factor Risk：
<mrow>
<mi>R</mi>
<mi>i</mi>
<mi>s</mi>
<mi>k</mi>
<mo>=</mo>
<mfrac>
<mrow>
<mn>1</mn>
<mo>-</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</msup>
</mrow>
<mrow>
<mn>1</mn>
<mo>+</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>t</mi>
<mo>)</mo>
</mrow>
</mrow>
</msup>
</mrow>
</mfrac>
<mo>&times;</mo>
<mn>100</mn>
<mi>%</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>9</mn>
<mo>)</mo>
</mrow>
</mrow>
If risk factor exceeds threshold value, send and confirm that short message confirms consumer behavior to user, and report supervision department to put on record.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710904426.3A CN107679982A (en) | 2017-09-29 | 2017-09-29 | A kind of credit card risk checking method based on point process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710904426.3A CN107679982A (en) | 2017-09-29 | 2017-09-29 | A kind of credit card risk checking method based on point process |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107679982A true CN107679982A (en) | 2018-02-09 |
Family
ID=61139287
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710904426.3A Pending CN107679982A (en) | 2017-09-29 | 2017-09-29 | A kind of credit card risk checking method based on point process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107679982A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135910A (en) * | 2019-05-16 | 2019-08-16 | 拉扎斯网络科技（上海）有限公司 | User data processing method and device, medium and computing equipment |
| CN110288157A (en) * | 2019-06-27 | 2019-09-27 | 电子科技大学 | A kind of Runoff Forecast method based on attention mechanism and LSTM |
| CN111161042A (en) * | 2019-11-26 | 2020-05-15 | 深圳壹账通智能科技有限公司 | Personal risk assessment method, device, terminal and storage medium |
| CN111461865A (en) * | 2020-03-31 | 2020-07-28 | 中国银行股份有限公司 | Data analysis method and device |
| CN111461866A (en) * | 2020-03-31 | 2020-07-28 | 中国银行股份有限公司 | Data analysis method and device |
| CN113505369A (en) * | 2021-09-08 | 2021-10-15 | 浙江大学 | Method and device for training user risk recognition model based on space-time perception |
-
2017
- 2017-09-29 CN CN201710904426.3A patent/CN107679982A/en active Pending
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135910A (en) * | 2019-05-16 | 2019-08-16 | 拉扎斯网络科技（上海）有限公司 | User data processing method and device, medium and computing equipment |
| CN110288157A (en) * | 2019-06-27 | 2019-09-27 | 电子科技大学 | A kind of Runoff Forecast method based on attention mechanism and LSTM |
| CN111161042A (en) * | 2019-11-26 | 2020-05-15 | 深圳壹账通智能科技有限公司 | Personal risk assessment method, device, terminal and storage medium |
| CN111461865A (en) * | 2020-03-31 | 2020-07-28 | 中国银行股份有限公司 | Data analysis method and device |
| CN111461866A (en) * | 2020-03-31 | 2020-07-28 | 中国银行股份有限公司 | Data analysis method and device |
| CN111461866B (en) * | 2020-03-31 | 2023-09-26 | 中国银行股份有限公司 | Data analysis method and device |
| CN111461865B (en) * | 2020-03-31 | 2024-02-02 | 中国银行股份有限公司 | Data analysis method and device |
| CN113505369A (en) * | 2021-09-08 | 2021-10-15 | 浙江大学 | Method and device for training user risk recognition model based on space-time perception |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107679982A (en) | A kind of credit card risk checking method based on point process | |
| CN109165950A (en) | A kind of abnormal transaction identification method based on financial time series feature, equipment and readable storage medium storing program for executing | |
| Choong et al. | Export-led growth hypothesis in Malaysia: An investigation using bounds test | |
| CN110376457A (en) | Non-intrusion type load monitoring method and device based on semi-supervised learning algorithm | |
| CA3066330A1 (en) | Improved detection of fraudulent transactions | |
| Narayan et al. | The relationship between the real exchange rate and balance of payments: empirical evidence for China from cointegration and causality testing | |
| US20200211020A1 (en) | Systems and methods for improved detection of network fraud events | |
| Tudela | Explaining currency crises: a duration model approach | |
| CN103678659A (en) | E-commerce website cheat user identification method and system based on random forest algorithm | |
| US20210312452A1 (en) | Systems and methods real-time institution analysis based on message traffic | |
| Jiang et al. | Short term prediction of extreme returns based on the recurrence interval analysis | |
| CN112365338A (en) | Artificial intelligence-based data fraud detection method, device, terminal and medium | |
| Gao et al. | Research on Default Prediction for Credit Card Users Based on XGBoost‐LSTM Model | |
| Tang et al. | A study of financial market resilience in China—From a hot money shock perspective | |
| Yang et al. | Multi-view broad learning system for electricity theft detection | |
| Oet et al. | From organization to activity in the US collateralized interbank market | |
| Falaki et al. | Probabilistic credit card fraud detection system in online transactions | |
| CN114331463A (en) | Risk identification method based on linear regression model and related equipment thereof | |
| Tsakonas et al. | Application of the moving Lyapunov exponent to the S&P 500 index to predict major declines | |
| Qianyu et al. | Research on financial consumer behavior based on deep Learning | |
| Mu | Credit card fraud intelligent detection based on machine learning | |
| Belbute | Is the Euro-Area core price index really more persistent than the food and energy price indexes? | |
| Tao et al. | Dynamic correlation and risk resonance among industries of Chinese stock market: New evidence from time–frequency domain and complex network perspectives | |
| Yu et al. | Construction and Application of Bid Fraud Prediction Model Based on AdaBoost Algorithm | |
| Baek et al. | “Left Shoulder” Detection in Korea Composite Stock Price Index Using an Auto-Associative Neural Network |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180209 |