CN109934700A - A kind of method and device of arbitrage detecting - Google Patents
A kind of method and device of arbitrage detecting Download PDFInfo
- Publication number
- CN109934700A CN109934700A CN201910154896.1A CN201910154896A CN109934700A CN 109934700 A CN109934700 A CN 109934700A CN 201910154896 A CN201910154896 A CN 201910154896A CN 109934700 A CN109934700 A CN 109934700A
- Authority
- CN
- China
- Prior art keywords
- arbitrage
- transaction
- trade
- credit card
- trade company
- 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
Abstract
The invention discloses a kind of method and devices of arbitrage detecting, this method includes, obtain the transaction data of credit card trade, according to the transaction data of credit card trade, arbitrage transaction Rating Model, determine the score value of credit card trade, if the score value of credit card trade is greater than scoring threshold value, then determine credit card trade for arbitrage transaction, wherein, arbitrage transaction Rating Model is the arbitrage label according to credit card trade, it is generated after first historical trading data progress Training, the arbitrage label of credit card trade is generated after carrying out unsupervised training according to the credit card trade of arbitrage trade company and each arbitrage trade company in arbitrage trade company pond.The technical solution improves the accuracy rate and coverage rate of arbitrage transaction detecting.
Description
Technical field
The present invention relates to the method and devices that technical field of data processing more particularly to a kind of arbitrage are detected.
Background technique
The risk of current credit card violation arbitrage is increasingly prominent, causes numerous adverse effects to society.Existing arbitrage transaction
Method for detecting construct Expert Rules mainly according to the transaction feature of card and trade company and identified.However, arbitrage method is got over
Come more hidden, technology is more and more stronger, and gimmick is more and more professional, and clique, specialized feature, and current arbitrage friendship is gradually presented
Easy sample data extremely lacks, and causes the accuracy rate of existing method detecting arbitrage transaction and coverage rate lower.
Summary of the invention
The embodiment of the present invention provides a kind of method and device of arbitrage detecting, to improve the accuracy rate of arbitrage transaction detecting
And coverage rate.
A kind of method of arbitrage detecting provided in an embodiment of the present invention, comprising:
Obtain the transaction data of credit card trade;
According to the transaction data of the credit card trade, arbitrage transaction Rating Model, commenting for the credit card trade is determined
Score value;Wherein, the arbitrage transaction Rating Model is carried out according to arbitrage label, the first historical trading data of credit card trade
It is generated after Training;The arbitrage label of the credit card trade is according to arbitrage trade company in arbitrage trade company pond and each
What the credit card trade of arbitrage trade company generate after unsupervised training;
If the score value of the credit card trade is greater than scoring threshold value, it is determined that the credit card trade is arbitrage transaction.
In above-mentioned technical proposal, the transaction data of credit card trade is obtained, and according to the transaction data of credit card trade, set
It now trades Rating Model, determines the score value of the credit card trade, if the score value of the credit card trade is greater than scoring threshold value,
Determine the credit card trade for arbitrage transaction.Wherein, arbitrage transaction Rating Model is according to the arbitrage label of credit card trade, the
It is generated after one historical trading data progress Training, the arbitrage label of credit card trade is according in arbitrage trade company pond
What the credit card trade of arbitrage trade company and each arbitrage trade company generate after unsupervised training.That is, this programme is being commented
Estimate credit card trade whether be arbitrage transaction when, use arbitrage transaction Rating Model be based on unsupervised training, have supervision instruct
Practice generation, wherein unsupervised trained utilization is figure calculation method, more quick update, extended network in association is propagated;
What Training used is the scorecard model returned based on traditional logic, promotes the computational efficiency in iterative process.The two
Series connection optimization is carried out in a model, until forming the closed-loop system for completely restraining, having self-learning capability, improves arbitrage transaction
The accuracy rate and coverage rate of detecting.
Optionally, the transaction data according to the credit card trade, arbitrage transaction Rating Model, determine the credit
Block the score value of transaction, comprising:
Obtain each input variable of the arbitrage transaction Rating Model;
According to each input variable of the transaction data of the credit card trade, arbitrage transaction Rating Model, institute is determined
State the corresponding value of each input variable of arbitrage transaction Rating Model;
The corresponding value of each input variable of arbitrage transaction Rating Model is input to the arbitrage transaction Rating Model,
Obtain the score value of the credit card trade.
In above-mentioned technical proposal, each input variable of arbitrage transaction Rating Model is obtained, and according to the friendship of credit card trade
Each input variable of easy data, arbitrage transaction Rating Model, determines the corresponding value of each input variable of arbitrage transaction Rating Model,
Then after the corresponding value of each input variable determined being input to arbitrage transaction Rating Model, the credit card trade is calculated
Score value.
Optionally, first historical trading data includes the number of deals of multiple credit card trades in preset time period
According to the corresponding card archives of transaction data, trade company's archives, the Authority Profile, region archives of, each credit card trade;
It is described according to the arbitrage label of credit card trade, the first historical trading data carry out Training after generate described in
Arbitrage transaction Rating Model, comprising:
The tag variable of the arbitrage transaction Rating Model is determined according to the arbitrage label of the credit card trade;
According to the transaction data of multiple credit card trades in the preset time period, the transaction data of each credit card trade
Corresponding card archives, trade company's archives, Authority Profile, region archives determine the input variable of the arbitrage transaction Rating Model;
It is traded the tag variable of Rating Model, the input variable of arbitrage transaction Rating Model according to the arbitrage, using patrolling
It collects regression algorithm and carries out Training, until determining the arbitrage transaction Rating Model.
In above-mentioned technical proposal, determine that the label of arbitrage transaction Rating Model becomes according to the arbitrage label of credit card trade
Amount, according to the corresponding card archives of the transaction data of the transaction data of multiple credit card trades, each credit card trade, trade company's shelves
Case, Authority Profile, region archives determine that the input variable of arbitrage transaction Rating Model is adopted according to tag variable, input variable
Training is carried out with logistic regression algorithm, until determining arbitrage transaction Rating Model.It, can be accurate using there is monitor model
Identify whether credit card trade is arbitrage transaction, improves the accuracy rate of arbitrage transaction detecting.
Optionally, the credit card trade of the arbitrage trade company according in arbitrage trade company pond and each arbitrage trade company carries out nothing
The arbitrage label of the credit card trade is generated after supervised training, comprising:
According to the credit card trade of arbitrage trade company and each arbitrage trade company in arbitrage trade company pond, arbitrage card is determined
Set, the transaction relationship between the set of arbitrage trade company and each arbitrage card and each arbitrage trade company;
According between the set of the arbitrage card, the set of arbitrage trade company and each arbitrage card and each arbitrage trade company
Transaction relationship generates the arbitrage label of the credit card trade after carrying out repeatedly association propagation.
In above-mentioned technical proposal, according to the credit card trade of arbitrage trade company and each arbitrage trade company in arbitrage trade company pond,
Determine the transaction relationship between the set of arbitrage card, the set of arbitrage trade company and each arbitrage card and each arbitrage trade company, and
Repeatedly association is carried out to propagate, it, can will credit card trade associated with arbitrage trade company, arbitrage card when association, which is propagated, to be terminated
It is determined as arbitrage transaction, and generates the arbitrage label of credit card trade.It, can be in exemplar data deficiencies using unsupervised training
In the case where, arbitrage transaction identification is carried out, the arbitrage label of credit card trade is generated, substantially reduces the operation of label data acquisition
Cost.In addition, generating the arbitrage label of credit card trade by unsupervised training, resource overhead is smaller, is suitable for practical application,
It is associated in communication process, in addition to artificial progress parameter setting is needed when cold start-up, generally by the output result of Training
It determines, more fitting business scenario, and reliability is high.
Optionally, it includes that arbitrage card is propagated and the propagation of arbitrage trade company that the association, which is propagated,;
It is described according to the set of the arbitrage card, the set of arbitrage trade company and each arbitrage card and each arbitrage trade company it
Between transaction relationship, carry out repeatedly association propagate, comprising:
For any time in the multiple association propagation, according to the arbitrage trade company, each arbitrage card and each set
Transaction relationship between existing trade company determines that newly-increased arbitrage card is placed in the set of the arbitrage card;According to the arbitrage
Transaction relationship between card, each arbitrage card and each arbitrage trade company determines that newly-increased arbitrage trade company is placed in the set
The set of existing trade company.
In above-mentioned technical proposal, it includes that the propagation of arbitrage card and arbitrage trade company propagate that association, which is propagated, can be according to arbitrage quotient
Transaction relationship between family, each arbitrage card and each arbitrage trade company carries out the propagation of arbitrage card, and according to arbitrage card, each set
Transaction relationship between existing card and each arbitrage trade company carries out the propagation of arbitrage trade company, when without newly-increased arbitrage card or without newly-increased arbitrage
After trade company, it can determine that association is propagated and terminate.Using figure calculation method, more quickly updated in association is propagated, extended network,
And the arbitrage label of credit card trade is generated by unsupervised training, resource overhead is smaller, is suitable for practical application, is not only restricted to
The sample size of arbitrage transaction.
Optionally, after the determination credit card trade is arbitrage transaction, further includes:
The corresponding trade company of arbitrage transaction is determined as storing after the arbitrage trade company into arbitrage trade company pond.
In above-mentioned technical proposal, after credit card trade is confirmed as arbitrage transaction, which can be traded corresponding
Trade company is determined as arbitrage trade company, and stores into arbitrage trade company pond, by summarize arbitrage trade Rating Model output as a result, from
And realize the closed loop study of Training and unsupervised training.That is, when being associated propagation, the collection of arbitrage trade company
Arbitrage trade company in conjunction can be determined by the output result of Training, more be bonded business scenario, and reliability is high.
Correspondingly, the embodiment of the invention also provides a kind of devices of arbitrage detecting, comprising:
Acquiring unit, for obtaining the transaction data of credit card trade;
Processing unit determines the letter for transaction data, the arbitrage transaction Rating Model according to the credit card trade
The score value traded with card;Wherein, arbitrage transaction Rating Model is arbitrage label according to credit card trade, the first history
It is generated after transaction data progress Training;The arbitrage label of the credit card trade is according to the set in arbitrage trade company pond
What the credit card trade of existing trade company and each arbitrage trade company generate after unsupervised training;If the scoring of the credit card trade
Value is greater than scoring threshold value, it is determined that the credit card trade is arbitrage transaction.
Optionally, the processing unit is specifically used for:
Obtain each input variable of the arbitrage transaction Rating Model;
According to each input variable of the transaction data of the credit card trade, arbitrage transaction Rating Model, institute is determined
State the corresponding value of each input variable of arbitrage transaction Rating Model;
The corresponding value of each input variable of arbitrage transaction Rating Model is input to the arbitrage transaction Rating Model,
Obtain the score value of the credit card trade.
Optionally, first historical trading data includes the number of deals of multiple credit card trades in preset time period
According to the corresponding card archives of transaction data, trade company's archives, the Authority Profile, region archives of, each credit card trade;
The processing unit is specifically used for:
The tag variable of the arbitrage transaction Rating Model is determined according to the arbitrage label of the credit card trade;
According to the transaction data of multiple credit card trades in the preset time period, the transaction data of each credit card trade
Corresponding card archives, trade company's archives, Authority Profile, region archives determine the input variable of the arbitrage transaction Rating Model;
It is traded the tag variable of Rating Model, the input variable of arbitrage transaction Rating Model according to the arbitrage, using patrolling
It collects regression algorithm and carries out Training, until determining the arbitrage transaction Rating Model.
Optionally, the processing unit is specifically used for:
According to the credit card trade of arbitrage trade company and each arbitrage trade company in arbitrage trade company pond, arbitrage card is determined
Set, the transaction relationship between the set of arbitrage trade company and each arbitrage card and each arbitrage trade company;
According between the set of the arbitrage card, the set of arbitrage trade company and each arbitrage card and each arbitrage trade company
Transaction relationship generates the arbitrage label of the credit card trade after carrying out repeatedly association propagation.
Optionally, it includes that arbitrage card is propagated and the propagation of arbitrage trade company that the association, which is propagated,;
The processing unit is specifically used for:
For any time in the multiple association propagation, according to the arbitrage trade company, each arbitrage card and each set
Transaction relationship between existing trade company determines that newly-increased arbitrage card is placed in the set of the arbitrage card;According to the arbitrage
Transaction relationship between card, each arbitrage card and each arbitrage trade company determines that newly-increased arbitrage trade company is placed in the set
The set of existing trade company.
Optionally, the processing unit is also used to:
After the determination credit card trade is arbitrage transaction, the corresponding trade company of arbitrage transaction is determined as
It is stored after the arbitrage trade company into arbitrage trade company pond.
Correspondingly, the embodiment of the invention also provides a kind of calculating equipment, comprising:
Memory, for storing program instruction;
Processor executes above-mentioned arbitrage according to the program of acquisition for calling the program instruction stored in the memory
The method of detecting.
Correspondingly, the embodiment of the invention also provides a kind of computer-readable non-volatile memory medium, including computer
Readable instruction, when computer is read and executes the computer-readable instruction, so that computer executes above-mentioned arbitrage detecting
Method.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly introduced, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill in field, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing.
Fig. 1 is a kind of schematic diagram of system architecture provided in an embodiment of the present invention;
Fig. 2 is a kind of flow diagram of the method for arbitrage detecting provided in an embodiment of the present invention;
Fig. 3 is a kind of flow diagram for generating arbitrage transaction Rating Model provided in an embodiment of the present invention;
Fig. 4 is a kind of flow diagram of determining credit card trade score value provided in an embodiment of the present invention;
Fig. 5 is the flow diagram of the method for another arbitrage detecting provided in an embodiment of the present invention;
Fig. 6 is a kind of structural schematic diagram of the device of arbitrage detecting provided in an embodiment of the present invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention make into
It is described in detail to one step, it is clear that described embodiments are only a part of the embodiments of the present invention, rather than whole implementation
Example.Based on the embodiments of the present invention, obtained by those of ordinary skill in the art without making creative efforts
All other embodiment, shall fall within the protection scope of the present invention.
Fig. 1 illustratively shows the system architecture that the method that the embodiment of the present invention provides arbitrage detecting is applicable in, this is
Framework of uniting can be server 100 and terminal 200;Wherein, terminal 200 transfers accounts, remits money, pays and refer to receive that user issues
After order, transaction request message is sent to server 100, whether server 100 judges the transaction by the transaction request message
For arbitrage transaction.Wherein, terminal 200 can be ATM (Automatic Teller Machine, ATM), POS
(Point of Sales, point-of-sale terminal) and smart phone, tablet computer, laptop etc..
Based on foregoing description, Fig. 2 illustratively shows a kind of method of arbitrage detecting provided in an embodiment of the present invention
Process, the device which can be detected by arbitrage execute, which can be located in server, can be the server.Such as
Shown in Fig. 2, which is specifically included:
Step 201, the transaction data of credit card trade is obtained.
Optionally, batch obtains the transaction data of the same day all credit card trades, for the friendship of each credit card trade
Easy data score.
Step 202, according to the transaction data of the credit card trade, arbitrage transaction Rating Model, the credit card is determined
The score value of transaction.
Wherein, arbitrage transaction Rating Model is carried out according to arbitrage label, the first historical trading data of credit card trade
It is generated after Training;The arbitrage label of credit card trade be according in arbitrage trade company pond arbitrage trade company and each arbitrage
What the credit card trade of trade company generate after unsupervised training.
The embodiment of the present invention provides a kind of specific implementation of arbitrage label for generating credit card trade.
According to the credit card trade of arbitrage trade company and each arbitrage trade company in arbitrage trade company pond, the collection of arbitrage card is determined
It closes, the transaction relationship between the set and each arbitrage card and each arbitrage trade company of arbitrage trade company, further according to the collection of arbitrage card
It closes, the transaction relationship between the set and each arbitrage card and each arbitrage trade company of arbitrage trade company, after carrying out repeatedly association propagation,
Generate the arbitrage label of credit card trade.Wherein, it includes that the propagation of arbitrage card and arbitrage trade company propagate that association, which is propagated, for multiple
Any time in association propagation, the propagation of arbitrage card is first carried out, specifically, according to arbitrage trade company, each arbitrage card and each arbitrage
Transaction relationship between trade company determines that newly-increased arbitrage card is placed in the set of arbitrage card;Then arbitrage trade company is carried out again
It propagates, specifically, determining newly-increased arbitrage according to the transaction relationship between arbitrage card, each arbitrage card and each arbitrage trade company
Trade company is placed in the set of arbitrage trade company.It can be set when there is no newly-increased arbitrage trade company and/or sets in the set of arbitrage trade company
There is no after newly-increased arbitrage card in the set of existing card, i.e., repeatedly association propagation terminates for judgement, and will be with arbitrage trade company, arbitrage
The associated credit card trade of card is determined as arbitrage transaction, and generates the arbitrage label of credit card trade.Certainly, pass is being determined
When connection propagates termination, more loose propagation termination condition can also be rule of thumb set.
It should be noted that needing to select initial arbitrage trade company from arbitrage trade company pond when being associated propagation for the first time.
Wherein, arbitrage trade company pond is mainly determined as the credit card trade of arbitrage transaction by arbitrage case library and arbitrage transaction Rating Model
Related trade company is constituted, and herein, arbitrage transaction Rating Model is determined as trade company involved in the credit card trade of arbitrage transaction
It specifically describes in the following embodiments.
Include multiple arbitrage seeds trade company in arbitrage trade company pond, is selected in multiple arbitrage seeds trade company from arbitrage trade company pond
When selecting initial arbitrage trade company, need the transaction data for all credit card trades for summarizing the end of month as foundation, wherein credit card
The transaction data of transaction may include the record Major key of of that month all credit card trades, transaction trade company number, transaction amount, transaction
The corresponding data of multiple variables such as card number, transaction channel, the partial content in the transaction data of credit card trade can be such as 1 institute of table
Show.According to the moon turnover of each arbitrage seed trade company in arbitrage trade company pond, the parameters such as card number of trading, from arbitrage trade company pond
In determine initial arbitrage trade company, herein, the relevant information of initial arbitrage trade company can be as shown in table 2.
The transaction data partial content signal of the of that month credit card trade of table 1
Variable | Major key | Trade company number | The amount of money | Card number | Transaction channel | …… |
This month transaction 1 | ||||||
This month transaction 2 | ||||||
…… |
The partial content of 2 arbitrage seed trade company of table is illustrated
Variable | This month transaction total amount | Of that month transaction credit card number | Whether arbitrage seed trade company | …… |
Trade company number | ||||
Trade company number | ||||
…… |
Initial arbitrage trade company is determined as initializing the arbitrage trade company in the set of arbitrage trade company, is being associated biography for the first time
Sowing time, wherein arbitrage trade company propagates to be associated in the of that month credit card that transaction occurred therewith according to initial arbitrage trade company, and leads to
Cross certain card propagation parameter setting filter out arbitrage card and by additions arbitrage card set in, the propagation of arbitrage card
To propagate ginseng in the of that month trade company that transaction occurred therewith, and by certain trade company according to the set associative of existing arbitrage card
Number setting filter out arbitrage trade company and by addition arbitrage trade company set.Further, it is repeatedly closed according to the above method
Connection is propagated, and is terminated until association is propagated.
In the embodiment of the present invention, according to the trade company of credit card trades all in time window and card transaction relationship and setting
Propagation parameter (card propagation parameter, trade company's propagation parameter) develop arbitrage trade company and arbitrage card network, therefore can be square
Just whether confirming trade company or card arbitrage according to the historical trading data of trade company and card.During each round association is propagated, according to
The historical trading relationship of card and trade company, progressive expansion arbitrage trade network, can excavate the arbitrage behavior of clique.
In the embodiment of the present invention, the first historical trading data may include multiple credit card trades in preset time period
The corresponding card archives of transaction data, trade company's archives, the Authority Profile, region archives of transaction data, each credit card trade.According to
Arbitrage label, the first historical trading data of credit card trade generate arbitrage transaction Rating Model, tool after carrying out Training
Body can flow chart as shown in Figure 3.
Step 301, the tag variable of arbitrage transaction Rating Model is determined according to the arbitrage label of credit card trade.
According to the arbitrage label of credit card trade, the label as the modeling sample in arbitrage transaction Rating Model is generated.
Step 302, according to the transaction of the transaction data, each credit card trade of multiple credit card trades in preset time period
The corresponding card archives of data, trade company's archives, Authority Profile, region archives determine the input variable of arbitrage transaction Rating Model.
Obtain transaction data and the relevant trade company side, region side, mechanism of multiple credit card trades in preset time period
Side data, and construct modeling sample.Specifically, can be the transaction data and phase of multiple credit card trades in preset time period
Initial modeling sample of the trade company side, region side, mechanism side data of pass as arbitrage transaction Rating Model.Wherein, when this is default
Between the transaction data of multiple credit card trades in section can be the friendship of all credit card trades of the processing in preset time period
Easy data.In the embodiment of the present invention, preset time period is past certain time period, and the length of preset time period can be by this field
Technical staff self-setting, generally one month according to actual needs, or a season, but be also required to adjust accordingly
It is associated with the period propagated.
In the embodiment of the present invention, the transaction data of multiple credit card trades in the preset time period includes credit card trade
Arm's length dealing and credit card trade arbitrage transaction, and, each of transaction data of multiple credit card trade credit
The transaction data of card transaction all may include the exchange hour of the credit card trade, loco, transaction amount, transaction card number, hand over
The data such as easy card sending mechanism number, transaction trade company number, transaction acquirer number, transaction details, transaction channel.
Then, can according to the transaction data of multiple credit card trades in the preset time period and relevant trade company side,
Card archives relevant to multiple credit card trades, trade company's archives, region archives and mechanism are established in region side, mechanism side data
Archives.
One, card archives are established
According to the transaction data of multiple credit card trades in the preset time period, to be related in multiple credit card trades
Each card establish card archives, that is to say, that can be according to initial modeling sample, in past preset time period
There is each card of credit card trade to establish card archives.Wherein, the card archives of any card specifically may include the card
The static data of piece, historical trading data, historical trading statistical data.Time-based statistical information, the statistics based on event
A plurality of types of data such as amount.
Wherein, the static data of card is card number;Historical trading data can be transaction record of the card in complete transaction
Data, such as exchange hour, loco, transaction amount;Historical trading statistical data can be the completed transaction of card
Statistical data, the statistical data two types including time-based statistical data and based on event, time-based statistical number
According to may include again in short-term statistical data and it is long when statistical data, statistical data refers to the card in past shorter one section in short-term
In (e.g., past 5 minutes, 10 minutes, 0.5 hour, 2 hours or 1 day) historical trading statistical data, card it is long when unite
It counts and refers to the card (e.g., past a couple of days, several weeks, several months or several years) historical trading within longer one end time in past
Statistical data, and, in short-term statistical data and it is long when statistical data all may include accumulation class, mean value class, accounting class, deviate class etc.
The corresponding data of a plurality of types of variables.By taking the statistical information in card one week as an example, accumulative category information may include that the card exists
Accumulation transaction total amount in past one week, variables, the mean value class such as accumulation transaction stroke count may include the card at the past one week
Interior average single transaction amount, accounting class may include the accounting that arbitrage is traded in all transaction in one week in the past, various types of
Accounting etc. of the transaction of type in all transaction;Statistical data based on event refers to that card is tired in past n transaction
A plurality of types of data, the n such as product class, mean value class, accounting class, deviation class can use multiple integer values, such as system of past 5 transaction
It counts, the statistical data of 10 transaction of past, the statistical data of past 50 transaction etc..
Two, trade company's archives are established
According to the transaction data of multiple credit card trades in the preset time period, to be related in multiple credit card trades
All trade companies establish trade company's archives, that is to say, that can be according to initial modeling sample, to have in past preset time period
Each trade company for crossing credit card trade establishes trade company's archives.Wherein, trade company's archives of any trade company, the trade company including the trade company
Number, a plurality of types of data such as historical trading data, historical trading statistical data.
Wherein, the unique designation number obtained when trade company number is Merchants register.Friendship can be completed for trade company in historical trading data
Easy transaction data, including but not limited to: exchange hour, transaction amount, transaction card number;Historical trading statistical data may include quotient
The statistical data of the completed transaction in family, the statistical data two types comprising time-based statistical data and based on event,
Time-based statistical data may include again in short-term statistical data and it is long when statistical data, statistical data refers to that the trade company exists in short-term
The statistics of (e.g., past 5 minutes, 10 minutes, 0.5 hour, 2 hours or 1 day) historical trading in shorter a period of time in past
Data, trade company it is long when statistical data refer to the trade company (e.g., past a couple of days, several weeks, number within long period of time in past
Month or the several years) historical trading statistical data, and in short-term statistical data and it is long when statistical data all may include accumulation class, mean value
Class, deviates the corresponding data of a plurality of types of variables such as class at accounting class.Including but not limited to: the trade company be averaged transaction amount with
The similar trade company of the whole industry is averaged transaction amount comparison, and store credit cards transaction stroke count accounts for the ratios of whole bank card business dealing stroke counts
Example.
Three, set up mechanism archives
According to the transaction data of multiple credit card trades in the preset time period, to be related in multiple credit card trades
All mechanism set up mechanism archives, that is to say, that according to initial modeling sample, to have letter in past preset time period
Each the acquirer set up mechanism archives traded with card.Wherein, the Authority Profile of any mechanism, the mechanism including the mechanism
Number, a plurality of types of data such as historical trading data, historical trading statistical data.
Wherein, mechanism number is mechanism mark unique code.Historical trading data can be the friendship of acquirer complete transaction
Easy data, including but not limited to: exchange hour, transaction amount, transaction trade company number, transaction card number;Historical trading statistical data can
Statistical data including the completed transaction of mechanism, two kinds of the statistical data comprising time-based statistical data and based on event
Type, time-based statistical data may include again in short-term statistical data and it is long when statistical data, statistical data refers to this in short-term
Mechanism (but longer compared to other files, e.g., past 1 day, 3 days or 1 week) historical trading within shorter a period of time in past
Statistical data, mechanism it is long when statistical data refer to the mechanism within longer one end time in past (e.g., past several weeks,
Several months, number season or the several years) historical trading statistical data, and in short-term statistical data and it is long when statistical data all may include accumulation
Class, accounting class, deviates the corresponding data of a plurality of types of variables such as class at mean value class.Including but not limited to: the mechanism averagely trades
Amount of money mechanism similar with the whole industry is averaged transaction amount comparison, which accounts for whole bank card business dealing amount of money
Ratio.
Four, region archives are established
According to the transaction data of multiple credit card trades in the preset time period, to be related in multiple credit card trades
All Activity area establish region archives, that is to say, that according to initial modeling sample, to have in past preset time period
Region archives are established in each the transaction area for crossing transaction.Wherein, the region archives of any area, the area including the region
Number, historical trading data, historical trading statistical data.
Wherein, region number is transaction geographical indicator unique code.Historical trading data can be regional complete transaction of trading
Transaction data, including but not limited to: exchange hour transaction amount, transaction trade company number, transaction card number;Historical trading statistical data
It may include the statistical data of the completed transaction in transaction area, comprising time-based statistical data and based on the statistics of event
Data two types, time-based statistical data may include again in short-term statistical data and it is long when statistical data, statistical number in short-term
According to referring to this area (but longer compared to other files, e.g., past 1 day, 3 days or 1 week) within shorter a period of time in past
The statistical data of historical trading, region it is long when statistical data refer to this area's (e.g., past within longer one end time in past
Several weeks, the several months, number season or the several years) historical trading statistical data, and in short-term statistical data and it is long when statistical data all may be used
Comprising accumulation class, mean value class, accounting class, deviate the corresponding data of a plurality of types of variables such as class.Including but not limited to: this area
The average average transaction amount comparison in transaction amount area similar with the whole industry, this area's credit card trade amount of money account for whole bank cards
The ratio of transaction amount, this area's night transaction accumulating sum.
Further, by the transaction of the transaction data, each credit card trade of multiple credit card trades in preset time period
The corresponding card archives of data, trade company's archives, Authority Profile, region archives form the modeling sample of arbitrage transaction Rating Model,
Include all changes in transaction data, relevant card archives, trade company's archives, Authority Profile and region archives in the modeling sample
Measure corresponding data.Can be according to the modeling sample of arbitrage transaction Rating Model, i.e., multiple credits card in preset time period are handed over
The corresponding card archives of transaction data, trade company's archives, the Authority Profile, region archives of easy transaction data, each credit card trade,
Determine the input variable of arbitrage transaction Rating Model.
In the embodiment of the present invention, can each of modeling sample variable first to arbitrage transaction Rating Model it is corresponding
Data carry out data cleansing and e.g. leave out or fill missing values therein and exceptional value.After carrying out data cleansing, trade to arbitrage
Each of modeling sample of Rating Model variable is grouped, and calculates the WOE of each variable by the following method
(Weight of Evidence, evidence weight) and IV (Information Value, the value of information).
Define the first variable, the first variable fingerstall now trade Rating Model modeling sample in any variable.By first
Variable corresponding data is divided into N group (N is positive integer).In the embodiment of the present invention, the packet mode and number of packet of each variable
It self-setting or can all be automated according to the actual situation using card side's branch mailbox, point of different variables by those skilled in the art
Group mode and number of packet can be identical, can not also be identical.
For the corresponding i-th group of data (1≤i≤N) of the first variable, i-th group of corresponding WOE is calculated, it can be such as formula (1)
It is shown:
Wherein, WOEiFor i-th group of WOE, the sample accounting of arbitrage transaction is that arbitrage is traded sample in i-th group in i-th group
Arbitrage is traded sample number in number/all groupings, in i-th group the sample accounting of arm's length dealing be arm's length dealing sample number in i-th group/
Arm's length dealing sample number in all groupings.
For the corresponding i-th group of data (1≤i≤N) of the first variable, i-th group of corresponding IV is calculated, it can be such as formula (2)
It is shown:
Wherein, IViFor i-th group of IV, WOEiFor i-th group of WOE.
After the corresponding WOE of N number of grouping of each variable is calculated, by the modeling sample of arbitrage transaction Rating Model
In the corresponding data of the variable replace with the WOE of grouping locating for the corresponding data of the variable, and then it is all variables are corresponding
Data all replace with the WOE value of respective packets, obtain the training sample of arbitrage transaction Rating Model.
After the corresponding IV of N number of grouping of each variable is calculated, the IV of all groupings of the first variable can be carried out
Summation, obtains the corresponding IV of the first variable, can be as shown in formula (3):
Since the IV of variable can reflect the predictive ability of the variable, IV value is bigger, illustrates the variable for judging a certain friendship
It whether is easily that the ability that arbitrage is traded is stronger, accordingly it is determined that the Rating Model that can trade from arbitrage being built out after the IV of each variable
M input variable of the biggish preceding M variable of IV as arbitrage transaction Rating Model is determined in all variables of apperance sheet.Its
In, M is positive integer.
Herein, it can be handed over according to the transaction data of multiple credit card trades in preset time period, and with multiple credits card
Easy relevant card archives, trade company's archives, Authority Profile and region archives, one or more for determining arbitrage transaction Rating Model
A input variable.
Step 303, it according to the input variable of the tag variable of arbitrage transaction Rating Model, arbitrage transaction Rating Model, adopts
Training is carried out with logistic regression algorithm, until determining arbitrage transaction Rating Model.
Optionally, logistic regression method can be used to be trained training sample, and use sensitivity analysis, correlation
The statistical methods such as inspection carry out supplemental training, determine arbitrage transaction Rating Model.
It, can also be using test sample to arbitrage after determining arbitrage transaction Rating Model in the embodiment of the present invention
Transaction Rating Model is tested, and after test passes through, and is used for arbitrage transaction Rating Model to assess credit card trade.This
Locate, the transaction data and correlation of multiple credit card trades in available the second preset time period before preset time period
Trade company side, region side, mechanism side data, and the friendship of multiple credit card trades in second preset time period got
Easy data and relevant trade company side, region side, mechanism side data are as initial testing sample.It then, can be second pre- according to this
If the transaction data of multiple credit card trades in the period and relevant trade company side, region side, mechanism side data, establish and more
The relevant card archives of a credit card trade, trade company's archives, region archives and Authority Profile, and test sample is formed, correspondingly,
Include all variables in transaction data, relevant card archives, trade company's archives, Authority Profile and region archives in test sample
Corresponding data.
Wherein, second preset time period is not be overlapped with preset time period, if preset time period is past 1 month, then
Second preset time can be one week before the preset time period, i.e., the sample in initial modeling sample and initial testing sample
Notebook data is not identical;And second preset time period size can by those skilled in the art's self-setting according to actual needs,
Preferably, the size of second preset time period can be equal to or less than the size of preset time period, that is to say, that initial testing
The quantity of sample in sample can be equal to or less than the sample size in initial modeling sample.
For example, as preset time period be past one month, the second preset time period be past two weeks, model is built
The vertical date is on September 1st, 2018, then initial modeling sample can be one of on August on August 31st, 1,1 2018
The transaction data of all credit card trades obtained in the time of the moon, wherein both including credit card arm's length dealing, also comprising believing
It is now traded with cutting ferrule;Correspondingly, initial testing sample can for two weeks of on July 31st, 18 days 1 July in 2018 when
In, the transaction data of all credit card trades of acquisition also includes credit card arbitrage wherein both including credit card arm's length dealing
Transaction.
In above-described embodiment, the implementation for determining arbitrage transaction Rating Model is elaborated, has determined this
Arbitrage is traded on the basis of Rating Model, can both be scored the credit card trade got, and determines that the credit card is handed over
It whether is easily arbitrage transaction.Specifically, being referred to the process shown in Fig. 4.
Step 401, each input variable of arbitrage transaction Rating Model is obtained.
Step 402, according to each input variable of the transaction data of credit card trade, arbitrage transaction Rating Model, set is determined
The now corresponding value of each input variable of transaction Rating Model.
In the embodiment of the present invention, card archives relevant to a certain credit card trade can be to carry out this credit card trade
Multiple variables corresponding transaction data of the card in card dimension, such as the day trade transaction total amount of card, the same day of card
Transaction stroke count etc..As shown in table 3, illustrate for the partial content of card archives.
The partial content of 3 card archives of table is illustrated
Variable | Day trade transaction total amount | Day trade transaction stroke count | …… |
Card mark |
Trade company's archives relevant to a certain credit card trade can be the trade company that is related in this credit card trade in quotient
The corresponding transaction data of multiple variables in the dimension of family, as certain trade company past one week in single be averaged transaction amount, certain trade company
Accounting etc. of the credit card trade carried out within past one week in all transaction.As shown in table 4, in the part of trade company's archives
Hold signal.
The partial content of 4 trade company's archives of table is illustrated
Authority Profile relevant to a credit card trade can be tieed up for the acquirer being related in the transaction in mechanism
The corresponding transaction data of multiple variables on degree, as certain acquirer past one week in credit card trade amount of money accounting, certain machine
The credit card that structure carried out within past one week pays out accounting etc. of the transaction in all transaction.It as shown in table 5, is Authority Profile
Partial content signal.
The partial content of 5 Authority Profile of table is illustrated
Region archives relevant to a credit card trade can be the transaction area that occurs in this credit card trade on ground
The corresponding transaction data of multiple variables in the dimension of domain, as somewhere past one week in credit card trade amount of money accounting, somewhere
Accounting etc. of the credit card night transaction that area carried out within past one week in all transaction.It as shown in table 6, is Authority Profile
Partial content signal.
The partial content of 6 region archives of table is illustrated
Step 403, the corresponding value of each input variable of arbitrage transaction Rating Model is input to arbitrage transaction Rating Model,
Obtain the score value of credit card trade.
It is interpreted as, it, will be each defeated after the corresponding value of each input variable that arbitrage transaction Rating Model is determined according to step 402
Enter the corresponding value of variable and be input to arbitrage transaction Rating Model, the corresponding value of each input variable herein can be each input variable
Corresponding score value, and export the score value of the credit card trade.According to the score value of the credit card trade and the threshold value that scores
Compare, it can determine whether this credit card trade is arbitrage transaction.
Step 203, if the score value of the credit card trade is greater than scoring threshold value, it is determined that the credit card trade is set
Now trade.
In step 202, after the score value of export credit card transaction, by the score value of the credit card trade and scoring threshold value
It makes comparisons, if the score value of credit card trade is greater than scoring threshold value, it is determined that this credit card trade is arbitrage transaction;If credit
The score value of card transaction is no more than scoring threshold value, it is determined that this credit card trade is arm's length dealing.
In the embodiment of the present invention, scoring threshold value, can be according to pre- for judging whether credit card trade is arbitrage transaction
If the transaction data of multiple credit card trades in the period, and card archives relevant to multiple credit card trades, trade company's shelves
Case, Authority Profile and region archives determine.Specifically, scoring threshold value can be determined by the external rules decision model of auxiliary,
Interpretation of result can also be propagated by association and determines scoring threshold value, or determines scoring threshold value by other means, do not do and have herein
Body limits.
As an example it is assumed that being 850 points according to the scoring threshold value that historical trading data is determined, if same day credit card trade
In certain credit card trade score value be 950 points, then can determine this credit card trade be arbitrage trade, if same day credit
The score value of certain credit card trade is 750 points in card transaction, then can determine that this credit card trade is non-arbitrage (normal)
Transaction.
In the embodiment of the present invention, it can be gone through according to the transaction data of historical tradings multiple in preset time period with multiple
The relevant card archives of history transaction, trade company's archives, Authority Profile and region archives building arbitrage transaction Rating Model, therefore,
It, can be according to card archives, trade company's archives, mechanism shelves when being scored by arbitrage transaction Rating Model same day credit card trade
Data in case and region archives, determine same day credit card trade score value, and judge whether same day credit card trade is arbitrage
Transaction detects the problem of omitting to can effectively avoid in credit card arbitrage transaction, improves the detecting effect of arbitrage transaction, control
Credit card arbitrage risk.
In addition, corresponding trade company that arbitrage can also be traded is determined as after determining that credit card trade is arbitrage transaction
It is stored after arbitrage trade company into arbitrage trade company pond.Namely in the arbitrage trade company pond that step 202 is mentioned mainly by arbitrage case library
And arbitrage transaction Rating Model is determined as that trade company involved in the credit card trade of arbitrage transaction is constituted.
Embodiment in order to preferably explain the present invention will describe the stream of arbitrage detecting under specific implement scene below
Journey, as shown in figure 5, specific as follows:
The flow chart is divided into three parts, respectively association medium process, arbitrage transaction Rating Model building process and arbitrage
Transaction Rating Model process for using.
One, it is associated with medium process
Association medium process is related to step 501~step 503, and association medium process is unsupervised training process, with figure
Calculation method, it is therefore an objective to more quickly be updated in association is propagated, extended network.
Step 501, initial arbitrage trade company and propagation parameter are determined.
Herein, initial arbitrage trade company is determined from the arbitrage seed trade company in arbitrage trade company pond.Propagation parameter can be this
Field technical staff self-setting according to actual needs.
Step 502, it is associated propagation.
According to the credit card trade of arbitrage trade company and each arbitrage trade company in arbitrage trade company pond, the collection of arbitrage card is determined
It closes, the transaction relationship between the set and each arbitrage card and each arbitrage trade company of arbitrage trade company;And according to the collection of arbitrage card
It closes, the transaction relationship between the set and each arbitrage card and each arbitrage trade company of arbitrage trade company, carries out repeatedly association and propagate.
Step 503, the arbitrage label of credit card trade is generated.
Two, arbitrage transaction Rating Model constructs process
Arbitrage transaction Rating Model building process is related to step 504~step 508, and arbitrage transaction Rating Model constructs process
For Training process, with the scorecard model returned based on traditional logic, it is therefore an objective to promote the calculating in iterative process
Efficiency.
Step 504, the input variable of arbitrage transaction Rating Model is determined.
The first historical trading data is obtained, which includes that multiple credits card in preset time period are handed over
The corresponding card archives of transaction data, trade company's archives, the Authority Profile, region archives of easy transaction data, each credit card trade,
The input variable of arbitrage transaction Rating Model is determined according to above-mentioned data.
Step 505, the tag variable of arbitrage transaction Rating Model is determined.
By getting the arbitrage label of credit card trade in step 503, is determined and covered according to the arbitrage label of credit card trade
The now tag variable of transaction Rating Model.
Step 506, arbitrage transaction Rating Model training.
Training sample is trained using logistic regression method, and is counted using sensitivity analysis, correlation test etc.
Method carries out supplemental training, determines arbitrage transaction Rating Model.
Step 507, arbitrage transaction Rating Model assessment.
Arbitrage transaction Rating Model is assessed using test sample, until current arbitrage transaction Rating Model is by commenting
After estimating, it is determined as final mask out.
Step 508, arbitrage transaction Rating Model is determined.
Three, arbitrage transaction Rating Model process for using
Arbitrage transaction Rating Model process for using is related to step 509~step 512, in 501~step 508 of above-mentioned steps
In, it has been determined that go out final arbitrage transaction Rating Model, which can be used for daily
The transaction data of the credit card trade got is assessed.In addition, in arbitrage transaction Rating Model process for using, once really
Certain credit card trade is made as arbitrage transaction, it can the trade company that the credit card trade is related to is determined as arbitrage trade company, and
It stores into arbitrage trade company pond, the association for being associated in medium process is propagated, that is to say, that arbitrage transaction Rating Model is defeated
Result out is input to again in association medium process, and association medium process is determined in a manner of data characteristics driving, is realized
Automation association is propagated, and whole system is to have the closed-loop system of self-learning capability.
Step 509, each input variable of arbitrage transaction Rating Model is obtained.
Herein, each input variable of arbitrage transaction Rating Model is the step of Rating Model building process is traded in arbitrage
The input variable for the arbitrage transaction Rating Model determined in 504.
Step 510, the corresponding value of each input variable of arbitrage transaction Rating Model is determined.
According to each input variable of the transaction data of credit card trade, arbitrage transaction Rating Model, determine that arbitrage transaction is commented
The corresponding value of each input variable of sub-model.
Step 511, the score value of credit card trade is calculated.
After the corresponding value of each input variable of arbitrage transaction Rating Model is input to arbitrage transaction Rating Model, calculate
The score value of the credit card trade.
Step 512, judge whether the score value of credit card trade is greater than scoring threshold value.
If so, determining that credit card trade for arbitrage transaction, that is, turns to step 513.Otherwise, it determines credit card trade is non-
Arbitrage is traded (arm's length dealing), without processing.
Step 513, credit card trade is arbitrage transaction.
Step 514, the arbitrage corresponding trade company that trades is determined as arbitrage trade company.
It should be noted that association medium process, arbitrage transaction Rating Model building process, arbitrage transaction Rating Model make
It can be three processes parallel simultaneously with process, that is, be associated with the arbitrage label that medium process constantly updates credit card trade, and
The arbitrage label of updated credit card trade is input to arbitrage transaction Rating Model building process, arbitrage transaction Rating Model
It constructs process and constantly updates Rating Model, and updated Rating Model is input to arbitrage transaction Rating Model process for using,
It scores for the transaction data to the credit card got, and the corresponding arbitrage trade company of the arbitrage determined transaction is stored
Into arbitrage trade company pond, for being associated with the update of medium process.It is also possible to credit card trade in setting association medium process
The update cycle of arbitrage label, after being updated in certain update cycle to the arbitrage label of credit card trade, after update
Credit card trade arbitrage label be input to arbitrage transaction Rating Model building process, while be arranged arbitrage transaction Rating Model
The update cycle for constructing process, after being updated in certain update cycle to arbitrage transaction Rating Model, by updated set
Now transaction Rating Model be input to arbitrage transaction Rating Model process for using, for the transaction data to the credit card got into
Row scoring, and the corresponding arbitrage trade company of the arbitrage determined transaction is stored into arbitrage trade company pond, for being associated with medium process
Next update cycle in the current update cycle be updated.
In above-mentioned technical proposal, the transaction data of credit card trade is obtained, and according to the transaction data of credit card trade, set
It now trades Rating Model, determines the score value of the credit card trade, if the score value of the credit card trade is greater than scoring threshold value,
Determine the credit card trade for arbitrage transaction.Wherein, arbitrage transaction Rating Model is according to the arbitrage label of credit card trade, the
It is generated after one historical trading data progress Training, the arbitrage label of credit card trade is according in arbitrage trade company pond
What the credit card trade of arbitrage trade company and each arbitrage trade company generate after unsupervised training.That is, this programme is being commented
Estimate credit card trade whether be arbitrage transaction when, use arbitrage transaction Rating Model be based on unsupervised training, have supervision instruct
Practice generation, the Training of unsupervised training and logistic regression scoring that the association based on nomography is propagated, two training
It is iterated study after series connection, can carry out arbitrage identification in the case where exemplar data deficiencies, substantially reduce label data
The operation cost of acquisition.And store after being summarized the output result of arbitrage transaction Rating Model into arbitrage trade company pond, it is real
Existing supervised training and the study of the closed loop of unsupervised training, improve the accuracy rate of arbitrage transaction detecting, can identify extensive
The arbitrage of clique's property, including arbitrage trade company, channel and mechanism.
Based on the same inventive concept, Fig. 6 illustratively shows a kind of dress of arbitrage detecting provided in an embodiment of the present invention
The structure set, the device can execute the process of the method for arbitrage detecting.
Acquiring unit 601, for obtaining the transaction data of credit card trade;
Processing unit 602 determines institute for transaction data, the arbitrage transaction Rating Model according to the credit card trade
State the score value of credit card trade;Wherein, arbitrage transaction Rating Model is according to the arbitrage label of credit card trade, first
It is generated after historical trading data progress Training;The arbitrage label of the credit card trade is according in arbitrage trade company pond
Arbitrage trade company and the credit card trade of each arbitrage trade company generate after unsupervised training;If the credit card trade
Score value is greater than scoring threshold value, it is determined that the credit card trade is arbitrage transaction.
Optionally, the processing unit 602 is specifically used for:
Obtain each input variable of the arbitrage transaction Rating Model;
According to each input variable of the transaction data of the credit card trade, arbitrage transaction Rating Model, institute is determined
State the corresponding value of each input variable of arbitrage transaction Rating Model;
The corresponding value of each input variable of arbitrage transaction Rating Model is input to the arbitrage transaction Rating Model,
Obtain the score value of the credit card trade.
Optionally, first historical trading data includes the number of deals of multiple credit card trades in preset time period
According to the corresponding card archives of transaction data, trade company's archives, the Authority Profile, region archives of, each credit card trade;
The processing unit 602 is specifically used for:
The tag variable of the arbitrage transaction Rating Model is determined according to the arbitrage label of the credit card trade;
According to the transaction data of multiple credit card trades in the preset time period, the transaction data of each credit card trade
Corresponding card archives, trade company's archives, Authority Profile, region archives determine the input variable of the arbitrage transaction Rating Model;
It is traded the tag variable of Rating Model, the input variable of arbitrage transaction Rating Model according to the arbitrage, using patrolling
It collects regression algorithm and carries out Training, until determining the arbitrage transaction Rating Model.
Optionally, the processing unit 602 is specifically used for:
According to the credit card trade of arbitrage trade company and each arbitrage trade company in arbitrage trade company pond, arbitrage card is determined
Set, the transaction relationship between the set of arbitrage trade company and each arbitrage card and each arbitrage trade company;
According between the set of the arbitrage card, the set of arbitrage trade company and each arbitrage card and each arbitrage trade company
Transaction relationship generates the arbitrage label of the credit card trade after carrying out repeatedly association propagation.
Optionally, it includes that arbitrage card is propagated and the propagation of arbitrage trade company that the association, which is propagated,;
The processing unit 602 is specifically used for:
For any time in the multiple association propagation, according to the arbitrage trade company, each arbitrage card and each set
Transaction relationship between existing trade company determines that newly-increased arbitrage card is placed in the set of the arbitrage card;According to the arbitrage
Transaction relationship between card, each arbitrage card and each arbitrage trade company determines that newly-increased arbitrage trade company is placed in the set
The set of existing trade company.
Optionally, the processing unit 602 is also used to:
After the determination credit card trade is arbitrage transaction, the corresponding trade company of arbitrage transaction is determined as
It is stored after the arbitrage trade company into arbitrage trade company pond.
Based on the same inventive concept, the embodiment of the invention also provides a kind of calculating equipment, comprising:
Memory, for storing program instruction;
Processor executes above-mentioned arbitrage according to the program of acquisition for calling the program instruction stored in the memory
The method of detecting.
Based on the same inventive concept, the embodiment of the invention also provides a kind of computer-readable non-volatile memory medium,
Including computer-readable instruction, when computer is read and executes the computer-readable instruction, so that computer execution is above-mentioned
The method of arbitrage detecting.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although preferred embodiments of the present invention have been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the scope of the invention.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (14)
1. a kind of method of arbitrage detecting characterized by comprising
Obtain the transaction data of credit card trade;
According to the transaction data of the credit card trade, arbitrage transaction Rating Model, the score value of the credit card trade is determined;
Wherein, the arbitrage transaction Rating Model is to have carried out prison according to arbitrage label, the first historical trading data of credit card trade
Supervise and instruct and is generated after practicing;The arbitrage label of the credit card trade be according in arbitrage trade company pond arbitrage trade company and each arbitrage
What the credit card trade of trade company generate after unsupervised training;
If the score value of the credit card trade is greater than scoring threshold value, it is determined that the credit card trade is arbitrage transaction.
2. the method as described in claim 1, which is characterized in that the transaction data according to the credit card trade, arbitrage
Transaction Rating Model, determines the score value of the credit card trade, comprising:
Obtain each input variable of the arbitrage transaction Rating Model;
According to each input variable of the transaction data of the credit card trade, arbitrage transaction Rating Model, the set is determined
The now corresponding value of each input variable of transaction Rating Model;
The corresponding value of each input variable of arbitrage transaction Rating Model is input to the arbitrage transaction Rating Model, is obtained
The score value of the credit card trade.
3. the method as described in claim 1, which is characterized in that first historical trading data includes in preset time period
The corresponding card archives of transaction data, the trade company's archives, mechanism shelves of the transaction data of multiple credit card trades, each credit card trade
Case, region archives;
Arbitrage label, first historical trading data according to credit card trade generates the arbitrage after carrying out Training
Transaction Rating Model, comprising:
The tag variable of the arbitrage transaction Rating Model is determined according to the arbitrage label of the credit card trade;
It is corresponding according to the transaction data of the transaction data of multiple credit card trades in the preset time period, each credit card trade
Card archives, trade company's archives, Authority Profile, region archives, determine the input variable of arbitrage transaction Rating Model;
According to the input variable of the tag variable of arbitrage transaction Rating Model, arbitrage transaction Rating Model, returned using logic
Reduction method carries out Training, until determining the arbitrage transaction Rating Model.
4. the method as described in claim 1, which is characterized in that the arbitrage trade company according in arbitrage trade company pond and each set
The credit card trade of existing trade company carries out the arbitrage label that the credit card trade is generated after unsupervised training, comprising:
According to the credit card trade of arbitrage trade company and each arbitrage trade company in arbitrage trade company pond, the collection of arbitrage card is determined
It closes, the transaction relationship between the set and each arbitrage card and each arbitrage trade company of arbitrage trade company;
According to the transaction between the set of the arbitrage card, the set of arbitrage trade company and each arbitrage card and each arbitrage trade company
Relationship generates the arbitrage label of the credit card trade after carrying out repeatedly association propagation.
5. method as claimed in claim 4, which is characterized in that it includes that arbitrage card is propagated and arbitrage trade company that the association, which is propagated,
It propagates;
It is described according between the set of the arbitrage card, the set of arbitrage trade company and each arbitrage card and each arbitrage trade company
Transaction relationship carries out repeatedly association and propagates, comprising:
For any time in the multiple association propagation, according to the arbitrage trade company, each arbitrage card and each arbitrage quotient
Transaction relationship between family determines that newly-increased arbitrage card is placed in the set of the arbitrage card;According to the arbitrage card,
Transaction relationship between each arbitrage card and each arbitrage trade company determines that newly-increased arbitrage trade company is placed in the arbitrage trade company
Set.
6. such as method described in any one of claim 1 to 5, which is characterized in that in the determination credit card trade be set
After now trading, further includes:
The corresponding trade company of arbitrage transaction is determined as storing after the arbitrage trade company into arbitrage trade company pond.
7. a kind of device of arbitrage detecting characterized by comprising
Acquiring unit, for obtaining the transaction data of credit card trade;
Processing unit determines the credit card for transaction data, the arbitrage transaction Rating Model according to the credit card trade
The score value of transaction;Wherein, the arbitrage transaction Rating Model is arbitrage label, the first historical trading according to credit card trade
It is generated after data progress Training;The arbitrage label of the credit card trade is according to the arbitrage quotient in arbitrage trade company pond
What the credit card trade of family and each arbitrage trade company generate after unsupervised training;If the score value of the credit card trade is big
In scoring threshold value, it is determined that the credit card trade is arbitrage transaction.
8. device as claimed in claim 7, which is characterized in that the processing unit is specifically used for:
Obtain each input variable of the arbitrage transaction Rating Model;
According to each input variable of the transaction data of the credit card trade, arbitrage transaction Rating Model, the set is determined
The now corresponding value of each input variable of transaction Rating Model;
The corresponding value of each input variable of arbitrage transaction Rating Model is input to the arbitrage transaction Rating Model, is obtained
The score value of the credit card trade.
9. device as claimed in claim 7, which is characterized in that first historical trading data includes in preset time period
The corresponding card archives of transaction data, the trade company's archives, mechanism shelves of the transaction data of multiple credit card trades, each credit card trade
Case, region archives;
The processing unit is specifically used for:
The tag variable of the arbitrage transaction Rating Model is determined according to the arbitrage label of the credit card trade;
It is corresponding according to the transaction data of the transaction data of multiple credit card trades in the preset time period, each credit card trade
Card archives, trade company's archives, Authority Profile, region archives, determine the input variable of arbitrage transaction Rating Model;
According to the input variable of the tag variable of arbitrage transaction Rating Model, arbitrage transaction Rating Model, returned using logic
Reduction method carries out Training, until determining the arbitrage transaction Rating Model.
10. device as claimed in claim 7, which is characterized in that the processing unit is specifically used for:
According to the credit card trade of arbitrage trade company and each arbitrage trade company in arbitrage trade company pond, the collection of arbitrage card is determined
It closes, the transaction relationship between the set and each arbitrage card and each arbitrage trade company of arbitrage trade company;
According to the transaction between the set of the arbitrage card, the set of arbitrage trade company and each arbitrage card and each arbitrage trade company
Relationship generates the arbitrage label of the credit card trade after carrying out repeatedly association propagation.
11. device as claimed in claim 10, which is characterized in that it includes that arbitrage card is propagated and arbitrage quotient that the association, which is propagated,
It propagates at family;
The processing unit is specifically used for:
For any time in the multiple association propagation, according to the arbitrage trade company, each arbitrage card and each arbitrage quotient
Transaction relationship between family determines that newly-increased arbitrage card is placed in the set of the arbitrage card;According to the arbitrage card,
Transaction relationship between each arbitrage card and each arbitrage trade company determines that newly-increased arbitrage trade company is placed in the arbitrage trade company
Set.
12. such as the described in any item devices of claim 7 to 11, which is characterized in that the processing unit is also used to:
After the determination credit card trade is arbitrage transaction, the corresponding trade company of arbitrage transaction is determined as described
It is stored after arbitrage trade company into arbitrage trade company pond.
13. a kind of calculating equipment characterized by comprising
Memory, for storing program instruction;
Processor requires 1 to 6 according to the program execution benefit of acquisition for calling the program instruction stored in the memory
Described in any item methods.
14. a kind of computer-readable non-volatile memory medium, which is characterized in that including computer-readable instruction, work as computer
When reading and executing the computer-readable instruction, so that computer executes such as method as claimed in any one of claims 1 to 6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910154896.1A CN109934700A (en) | 2019-03-01 | 2019-03-01 | A kind of method and device of arbitrage detecting |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910154896.1A CN109934700A (en) | 2019-03-01 | 2019-03-01 | A kind of method and device of arbitrage detecting |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109934700A true CN109934700A (en) | 2019-06-25 |
Family
ID=66986269
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910154896.1A Pending CN109934700A (en) | 2019-03-01 | 2019-03-01 | A kind of method and device of arbitrage detecting |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109934700A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110942312A (en) * | 2019-11-29 | 2020-03-31 | 智器云南京信息科技有限公司 | POS machine cash register identification method, system, equipment and storage medium |
CN112017029A (en) * | 2020-08-31 | 2020-12-01 | 中国银行股份有限公司 | Information prompting method and device |
CN112926991A (en) * | 2021-03-30 | 2021-06-08 | 顶象科技有限公司 | Cascade group severity grade dividing method and system |
CN112990919A (en) * | 2019-12-17 | 2021-06-18 | 中国银联股份有限公司 | Information processing method and device |
CN112926991B (en) * | 2021-03-30 | 2024-04-30 | 中国银联股份有限公司 | Method and system for grading severity level of cash-out group |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104636912A (en) * | 2015-02-13 | 2015-05-20 | 银联智惠信息服务(上海)有限公司 | Identification method and device for withdrawal of credit cards |
CN106447333A (en) * | 2016-11-29 | 2017-02-22 | 中国银联股份有限公司 | Fraudulent trading detection method and server |
WO2017032056A1 (en) * | 2015-08-26 | 2017-03-02 | 中兴通讯股份有限公司 | Point-of-sale-based cash-out determining method and apparatus |
CN109034209A (en) * | 2018-07-03 | 2018-12-18 | 阿里巴巴集团控股有限公司 | The training method and device of the real-time identification model of active risk |
-
2019
- 2019-03-01 CN CN201910154896.1A patent/CN109934700A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104636912A (en) * | 2015-02-13 | 2015-05-20 | 银联智惠信息服务(上海)有限公司 | Identification method and device for withdrawal of credit cards |
WO2017032056A1 (en) * | 2015-08-26 | 2017-03-02 | 中兴通讯股份有限公司 | Point-of-sale-based cash-out determining method and apparatus |
CN106447333A (en) * | 2016-11-29 | 2017-02-22 | 中国银联股份有限公司 | Fraudulent trading detection method and server |
CN109034209A (en) * | 2018-07-03 | 2018-12-18 | 阿里巴巴集团控股有限公司 | The training method and device of the real-time identification model of active risk |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110942312A (en) * | 2019-11-29 | 2020-03-31 | 智器云南京信息科技有限公司 | POS machine cash register identification method, system, equipment and storage medium |
CN112990919A (en) * | 2019-12-17 | 2021-06-18 | 中国银联股份有限公司 | Information processing method and device |
CN112017029A (en) * | 2020-08-31 | 2020-12-01 | 中国银行股份有限公司 | Information prompting method and device |
CN112017029B (en) * | 2020-08-31 | 2023-09-08 | 中国银行股份有限公司 | Information prompting method and device |
CN112926991A (en) * | 2021-03-30 | 2021-06-08 | 顶象科技有限公司 | Cascade group severity grade dividing method and system |
CN112926991B (en) * | 2021-03-30 | 2024-04-30 | 中国银联股份有限公司 | Method and system for grading severity level of cash-out group |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7165043B2 (en) | Valuation prediction models in situations with missing inputs | |
US8209251B2 (en) | Cross correlation tool for automated portfolio descriptive statistics | |
US7096197B2 (en) | Methods and apparatus for simulating competitive bidding yield | |
US7031936B2 (en) | Methods and systems for automated inferred valuation of credit scoring | |
KR100766149B1 (en) | Methods and systems for efficiently sampling portfolios for optimal underwriting | |
US7028005B2 (en) | Methods and systems for finding value and reducing risk | |
US6985881B2 (en) | Methods and apparatus for automated underwriting of segmentable portfolio assets | |
Řezáč et al. | How to measure the quality of credit scoring models | |
US7120599B2 (en) | Methods and systems for modeling using classification and regression trees | |
KR100796058B1 (en) | Methods and apparatus for rapid deployment of a valuation system | |
US7082411B2 (en) | Methods and systems for optimizing return and present value | |
KR100766148B1 (en) | Methods and systems for quantifying cash flow recovery and risk | |
US20020013752A1 (en) | Rapid valuation of portfolios of assets such as financial instruments | |
CN108053318A (en) | It is a kind of to the method and device that is identified of merchandising extremely | |
CN109934700A (en) | A kind of method and device of arbitrage detecting | |
CN109242279A (en) | A kind of appraisal procedure and server | |
CN110457159A (en) | A kind of method, apparatus, calculating equipment and the storage medium of processing batch tasks | |
CN109493086A (en) | A kind of method and device of determining violation trade company | |
CN107480703A (en) | Transaction fault detection method and device | |
CN108921607A (en) | A kind of Retail commodity Dynamic Pricing method, apparatus, equipment, system and medium | |
CN112037049B (en) | Bank card charging method and device | |
CN116308370A (en) | Training method of abnormal transaction recognition model, abnormal transaction recognition method and device | |
Lu et al. | Demand Deposits: Valuation and Interest Rate Risk Management | |
CN114549209A (en) | Intelligent risk assessment method, device, equipment and medium | |
Mayo | Cartesian genetic programming for trading: a preliminary investigation |
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 |