Specific embodiment
In order to better understand the above technical scheme, below by attached drawing and specific embodiment to this specification embodimentTechnical solution be described in detail, it should be understood that the specific features in this specification embodiment and embodiment are to this explanationThe detailed description of book embodiment technical solution, rather than the restriction to this specification technical solution, in the absence of conflict,Technical characteristic in this specification embodiment and embodiment can be combined with each other.
In a first aspect, this specification embodiment provides a kind of risk control method, as shown in Figure 1, implementing for this specificationThe flow chart for the risk control method that example provides, method includes the following steps:
Step S11: the target data of target trade company is obtained；
In this specification embodiment, target trade company can be any trade company for needing to carry out risk control, for example, target quotientFamily can be bank, service provider, self-employed businessman etc..Target data can be the data obtained in a predetermined time, for example, obtainingThe intraday data of target trade company are taken, or obtain the data etc. in 12 decimal of target trade company.Target data can be target trade companyThe data generated when operation are also possible to the build-in attribute data of target trade company, such as enterprise's license of target trade company, enterprise natureDeng.Target data can be used for carrying out the risk assessment of target trade company, and in one embodiment, target data includes multiple typesData, such as the transaction data etc. that trade company's attribute data, trade company generate, here without limitation.
Step S12: according to the target data and default risk score model, the risk of the target trade company is obtainedHealth data；
In this specification embodiment, default risk score model can be selected according to actual needs, be implemented at oneIn example, presetting risk score model is two disaggregated models, and in another embodiment, presetting risk score model is random forestModel.Default risk score model can be the preparatory model built using training sample and test sample.
It should be understood that the input of default risk score model may include various ways, for example, default risk score mouldThe input of type can be target data, be also possible to the feature extracted according to target data.The output of default risk score modelFor the risk health data of target trade company.In one embodiment, risk health data can be risk health score value.
Step S13: determining the corresponding target data ranges of the risk health data, and according to the target data modelCorresponding resource allocation rule is enclosed, distributes service resources for the target trade company.
In this specification embodiment, service resources can be the service resources that target trade company uses business, be also possible to meshMark trade company is not used but the service resources of business relevant to the business scope of target trade company.Not according to risk health dataTogether, the assigned degree of service resources can also be different.In this specification embodiment, risk health data can be divided into differenceGear, for example, by risk health data be risk health score value for, risk health score value is classified as within the scope of 0-60First grade, risk health score value is classified as second gear within the scope of 60-90, by risk health score value within the scope of 90-100It is classified as third gear.For different gears, different resource allocation rules can be set, above example is continued to use, first gradeResource allocation rule is to cancel allocated service resources, and the resource allocation rule of second gear is to subtract under absorbed service resources,The resource allocation rule of third gear is the service resources for increasing distribution.
In one embodiment, target trade company is used Alipay and receives single business, and the service resources of the business can beThe resource that returns the benefit of single business is received, for example, receiving single total value when target trade company carries out receiving single using the single business of Alipay receipts and being greater than oneWhen threshold value, disposably return the benefit 500 yuan to target trade company.If the risk health data of target trade company is higher, it can increaseReturn the benefit the amount of money, is encouraged and is affirmed come the risk health status current to target trade company.If the risk health of target trade companyWhen data are lower, then it can cancel and return the benefit to target trade company, to remind target trade company to be rectified and improved, control risk.
It in another embodiment, can be according to the number such as business scope of business information, target trade company that target trade company usesAccording to determining business relevant to the business scope of target trade company.For example, the business of target trade company is related to convenience-for-people payment, movementPayment, POS receive list, can determine the business scope of target trade company to receive single business.Receiving single business due to Alipay is and targetTherefore the relevant business of the business scope of trade company is not used Alipay in target trade company and receives single business, and the risk of target trade companyWhen health data is higher, the service resources for exempting from the single business of Alipay receipts of a month rate can be provided to target trade company, withThe risk health status current to target trade company is encouraged and is affirmed.
It is of course also possible to use other traffic resource assignment modes, target trade company such as high to risk health data increaseAdd traffic resource assignment, the service resources of the target trade company low to risk health data are not processed, here without limitation.
Optionally, the target data for obtaining target trade company, comprising: obtain the target trade company under target dimensionData are as the target data, and the target dimension includes one or more of following dimension: trade company's attribute dimensions, trade companyValue relevance, trade company's Social behaviors dimension and trade company's Risk Dimensions.
In this specification embodiment, in order to which the risk to target trade company is identified comprehensively, target data may include moreData under a dimension.Wherein, the corresponding data of trade company's attribute dimensions may include trade company's qualification, trade company's industry and commerce complete dataProperty, trade company whether Concern Mafia etc..The corresponding data of trade company's value relevance may include Merchants register duration, trade company's transaction data, quotientFamily trade user data etc..The corresponding data of trade company's Social behaviors dimension may include the relationship network data of trade company, social evaluationData etc..The corresponding data of trade company's Risk Dimensions may include merchant fraud risk trade data, trade company's arbitrage risk trade numberAccording to, trade company's wash sale risk trade data, trade company's gambling risk trade data, trade company's public sentiment risk trade data etc..
Optionally, the default risk score model is obtained by following manner: obtaining sample data sets, the sampleData acquisition system includes black sample data and white sample data, and the black sample data is that there are risk record trade companies to correspond to for historyData, the white sample data be history be not present the corresponding data of risk record trade company；According to the sample data sets,Obtain the default risk score model.
It may include training sample, test sample in sample data sets in this specification embodiment, wherein training sampleThis is used to examine the degree of reliability for training the model come for carrying out model training, test sample.It can in sample data setsTo be made of black sample data and white sample data, wherein there are the trade companies of risk record as black sample, the trade company for historyHistorical risk data can be used as one group of black sample data, there is no the trade companies of risk record as white sample, the quotient for historyThe historical data at family can be used as one group of white sample data.It should be noted that historical record can be the record in preset time,For example, using the record in three months as historical record.
It should be understood that may include the corresponding black sample data of multiple black sample trade companies in sample data sets, andThe corresponding white sample data of multiple white sample trade companies.It, can be with for sample data black for one group or one group of white sample dataData including the trade company under multiple dimensions, for example, one group of black sample data may include that the black sample trade company belongs in trade companyData under property dimension, trade company's value relevance, trade company's Social behaviors dimension and trade company's Risk Dimensions.
After getting sample data sets, model is trained by sample data sets.In one embodimentIn, after getting sample data sets, Characterizations can be carried out to every group of sample data, obtain data characteristics, thenModel training is carried out by data characteristics.Model training directly can also be carried out using sample data, here without limitation.TrainingModel can be selected according to actual needs, in one embodiment, model be two disaggregated models, wherein two classificationThe output of model is the probability that input sample is white sample, can be using this probability value as the risk health number of input sampleAccording to.For example, using one group of sample data of a sample as the input of two disaggregated models, if the output of two disaggregated models should notSample is that the probability of white sample is 98%, then the risk health data of the sample is 98 points.It is, of course, also possible to using other modesDefine risk health data, here without limitation.
Optionally, described according to the target data and default risk score model, obtain the wind of the target trade companyDangerous health data, comprising: according to the target data, determine the target trade company there is currently N kind risk classifications, N is positiveInteger；Risk data corresponding with every kind of risk classifications in the N kind risk classifications is determined in the target data；RootAccording to the risk data, single risk health data corresponding with every kind of risk classifications is determined, amount to and obtain N number of single windDangerous health data；According to N number of single risk health data and the default risk score model, the target is obtainedThe risk health data of trade company.
In this specification embodiment, since target data has included the data under multiple dimensions, target data can be withData comprising assessing multiple risk classifications.For example, the target data of target trade company includes that the relationship network data of trade company, social activity are commentedValence mumber evidence, trade company's transaction data, trade company's trade user data, wherein the relationship network data of trade company and social evaluation data canTo be used to determine the public sentiment risk of target trade company, trade company's transaction data and trade company's trade user data may be used to determine targetThe financial risks of trade company, i.e. target trade company there is currently risk classifications be two kinds, N 2.
Further, according to the risk data under every kind of risk classifications, the single risk to determine every kind of risk classifications is strongHealth data.The single risk health data of every kind of risk classifications can also be calculated according to model, in one embodiment, needleTo each risk classifications, one model of training can be corresponded to, the input of model can be the risk number under the risk classificationsAccording to the output of model is the corresponding single risk health data of the risk classifications.Certainly, the single risk of every kind of risk classifications is strongHealth data can also determine according to other modes, here without limitation.
After getting the single risk health data of every kind of risk classifications, can according to default risk score model,Obtain the risk health data of the target trade company.In this embodiment, the input for presetting risk score model is every kind of windThe single risk health data of dangerous type, the output for presetting risk score model is the risk health data of target trade company.Certainly,The single risk health data of every kind of risk classifications can also be obtained to the risk of target trade company by modes such as weighting processingHealth data, here without limitation.
Optionally, as shown in Fig. 2, the flow chart of the Risk Identification Method provided for this specification embodiment, this method is alsoInclude:
Step S21: according to the target data, determine the target trade company there is currently M kind risk classifications, M is positiveInteger；
Step S22: the determining risk identification with every kind of risk classifications in the M kind risk classifications obtains M as a result, amounting toKind risk identification result；
Step S23: being sent to the target trade company for the M kind risk identification result so that the target trade company according toThe M kind risk identification result takes corresponding risk processing operation.
In this specification embodiment, since target data has included the data under multiple dimensions, target data can be withData comprising assessing multiple risk classifications.For example, the target data of target trade company includes that the relationship network data of trade company, social activity are commentedValence mumber evidence, trade company's transaction data, trade company's trade user data, wherein the relationship network data of trade company and social evaluation data canTo be used to determine the public sentiment risk of target trade company, trade company's transaction data and trade company's trade user data may be used to determine targetThe financial risks of trade company, i.e. target trade company there is currently risk classifications be two kinds, M 2.
Further, the risk identification of every kind of risk classifications is determined as a result, in one embodiment, it can be according to every kind of riskType corresponding risk data determines risk identification result.For example, can be used according to trade company's transaction data and trade company's transactionUser data determines the financial risks recognition result of target trade company, it should be understood that may include in financial risks recognition resultRisk record of many aspects, such as risk of fraud record, gambling risk record, cheating risk record etc..Trade company's transaction dataAnd trade company's trade user data can carry risky label, e.g., in the transaction data of target trade company include 20 transactionData have risk of fraud label, then the financial risks recognition result of the target trade company is to include 20 risk of fraud notesRecord.
After obtaining the risk identification result of various risk classifications, risk identification result can be showed to target trade company,As shown in figure 3, showing the schematic diagram of the risk assessment display interface of target trade company for this specification embodiment.In the embodimentIn, risk assessment display interface can be shown on the electronic equipment of target trade company, such as mobile phone, apparatus such as computer.Target dataFor data of the collected target trade company within the time.The risk health data of target user is determined according to target data,And it will be shown on display interface, in this embodiment, risk health data is risk health score value, as shown in figure 3, the meshThe health risk score value for marking trade company is 95 points.According to the target data of target trade company, determine target trade company there are financial risks,The risk of these three types of account risk and public sentiment risk, i.e. M are 3.Wherein, financial risks recognition result to be taken advantage of for there are 20Cheat risk record, 15 gambling risk records, 2 cheating risk records.Account risk identification result is that there are 1 fund exceptionsExpenditure record.Public sentiment risk identification result is that there are 1 public sentiment risk records.
In addition, this specification embodiment can also provide key risk processing, i.e., as shown in figure 3, being directed to existing riskThe horse back that every kind of risk in Fig. 3 is shown below handles key.After detecting that user clicks the key, then risk is executed automaticallyProcessing, certainly, the record for needing target trade company voluntarily to handle or checked can remind target trade company voluntarily to execute.
Optionally, described according to the risk health data, service resources are distributed for the target trade company, comprising: according toThe historical risk health data of the target trade company and the risk health data determine that the risk of the target trade company is strongHealth data and curves；The maximum value of the risk health data curve is small meet first object data area when, be reduced to describedThe service resources of target trade company distribution, to remind the target trade company to handle risk；It is bent in the risk health dataWhen the minimum value of line meets the second target data ranges, increase the service resources for target trade company distribution.
In this specification embodiment, the risk health data of target trade company can be calculated primary and be pushed away every a preset timeGive target trade company.By taking risk health data is risk health score value as an example, the risk health to target trade company it can divide dailyValue is calculated, and the risk health score value on the same day is pushed to target trade company.If should be understood that the risk of target trade companyHealthy score value shows that the risk control of target trade company is preferable within a certain period of time if all maintaining higher score value.IfThe risk health score value of target trade company persistently drops within a certain period of time, then it is very big hidden to show that the risk of target trade company existsSuffer from, and target trade company is not rectified and improved.If the risk health score value of target trade company starts lower, maintained later higherScore value then shows that target trade company is rectified and improved for existing risk, and maintains the level after rectification.
Therefore, it in this specification embodiment, according to the historical risk health data of target trade company and currently getsRisk health data, draw target trade company risk health data curve, then according to the tendency of risk health data curve comeDetermine target trade company to the control situation of risk.Historical risk health data can be set according to actual needs, for example, going throughHistory risk health data is the risk health data calculated daily in the previous moon, or is every other day counted in first trimesterThe risk health data of calculation, here without limitation.
In this specification embodiment, mesh can be distinguished according to first object data area and the second target data rangesThe risk variation of trade company is marked, first object data area and the second target data ranges can be set according to actual needsIt is fixed.In one embodiment, first object data area be less than or equal to first threshold, the second target data ranges be greater thanOr it is equal to second threshold, and first threshold is less than or equal to second threshold.So in this embodiment, when risk health data is bentWhen the maximum value of line is less than or equal to first threshold, the items of measures to rectify and reform is not taken there are risk and corresponding to target trade companyPart.When the minimum value of risk health data curve is greater than or equal to second threshold, it is always maintained at corresponding to target trade company lowerThe condition of risk.
It is, of course, also possible to include other risk situations of change, e.g., when risk health data curve shows a increasing trend, and windDangerous health data finally keeps stable state over time, and there are small for risk health data when unsteady stateIn or equal to the score value of first threshold, stable state risk health data when being greater than or equal to second threshold, correspond to targetThere are risk and the conditions rectified and improved to risk for trade company.First threshold and second threshold can be selected according to actual needsIt selects, here without limitation.
In this specification embodiment, in order to encourage the higher target trade company of risk health data to continue to keep, it can increaseFor the service resources of target trade company distribution, in order to which the target trade company for alerting there are risk is rectified and improved, it is possible to reduce be target quotientThe service resources of family distribution.In addition, when encouraging target trade company, when can be according to the maintenance of high risk health dataBetween length Encourage the mechanism is classified, can also be according to risk health data likewise, when alerting target trade companyHeight carry out the classification of warning degree.
In one embodiment, service resources are the resource that returns the benefit for receiving single business, which is to make in target serviceWhen reaching a threshold value with the receipts list total value that Alipay receives single business, disposably return the benefit 500 yuan.When the risk health number of target trade companyWhen being always maintained at higher healthy score value according to curve table improving eyesight mark trade company, the resource that can will return the benefit, which increases to, disposably returns the benefit 800Member is affirmed with the risk control to target trade company.When the risk health data curve table improving eyesight mark trade company pair of target trade companyExisting risk is rectified and improved, and keeps risk health data when being in higher score value after rectifying and improving, and can will return the benefit resourceIncrease to and disposably return the benefit 600 yuan, target trade company is encouraged with appropriate, and target trade company is encouraged to continue to keep.Work as meshThe risk health data curve table improving eyesight mark trade company of mark trade company does not take there are risk and measures to rectify and reform, can cancel the money that returns the benefitSource, or close and receive single business, to remind target trade company to take measures to rectify and reform.
It should be understood that volume is, for the different situations of different risk health data curves, service resources can be pre-setThe allocation rule of distribution, when risk health data curve reflects corresponding situation, in the case of can calling directly this kindTraffic resource assignment rule.
Second aspect, this specification embodiment provides a kind of risk control device, referring to FIG. 4, including:
Data acquisition module 41, for obtaining the target data of target trade company；
Risk health data obtains module 42, for obtaining according to the target data and default risk score modelThe risk health data of the target trade company；
Resource distribution module 43, for being adjusted to the business rule of target service according to the risk health data,Wherein, the target service is and the associated business of target trade company.
In a kind of optional implementation, data acquisition module 41 is used for:
Data of the target trade company under target dimension are obtained as the target data, the target dimension include withOne or more of lower dimension: trade company's attribute dimensions, trade company's value relevance, trade company's Social behaviors dimension and trade company's risk dimensionDegree.
In a kind of optional implementation, described device further include:
Sample acquisition module, for obtaining sample data sets, the sample data sets include black sample data andWhite sample data, the black sample data are history there are the corresponding data of risk record trade company, and the white sample data is to go throughThe corresponding data of risk record trade company are not present in history；
Model obtains module, for obtaining the default risk score model according to the sample data sets.
In a kind of optional implementation, risk health data obtains module 42 and is used for:
According to the target data, determine the target trade company there is currently N kind risk classifications, N is positive integer；
Risk number corresponding with every kind of risk classifications in the N kind risk classifications is determined in the target dataAccording to；
According to the risk data, single risk health data corresponding with every kind of risk classifications is determined, it is total to obtainObtain N number of single risk health data；
According to N number of single risk health data and the default risk score model, the target trade company is obtainedRisk health data.
In a kind of optional implementation, described device further include:
Risk classifications determining module, for according to the target data, determine the target trade company there is currently M kind windDangerous type, M are positive integer；
Risk identification module, for the determining risk identification with every kind of risk classifications in the M kind risk classifications as a result,It is total to obtain M kind risk identification result；
Sending module, for the M kind risk identification result to be sent to the target trade company, so that the target trade companyCorresponding risk processing operation is taken according to the M kind risk identification result.
In a kind of optional implementation, resource distribution module 43 is used for:
According to the historical risk health data of the target trade company and the risk health data, the target is determinedThe risk health data curve of trade company；
When the maximum value of the risk health data curve meets first object data area, it is reduced to the target quotientThe service resources of family distribution, to remind the target trade company to handle risk；
When the minimum value of the risk health data curve meets the second target data ranges, increase as the target quotientThe service resources of family distribution.
About above-mentioned apparatus, wherein the concrete function of modules is in risk control side provided in an embodiment of the present inventionIt is described in detail in the embodiment of method, no detailed explanation will be given here.
The third aspect is based on inventive concept same as previous embodiment risk control method, and the present invention also provides oneKind of server is set, as shown in figure 5, including memory 604, processor 602 and being stored on memory 604 and can be in processorThe computer program run on 602, the processor 602 realize previously described risk control method when executing described programThe step of either method.
Wherein, in Fig. 5, bus architecture (is represented) with bus 600, and bus 600 may include any number of interconnectionBus and bridge, bus 600 will include the one or more processors represented by processor 602 and what memory 604 represented depositsThe various circuits of reservoir link together.Bus 600 can also will peripheral equipment, voltage-stablizer and management circuit etc. itVarious other circuits of class link together, and these are all it is known in the art, therefore, no longer carry out further to it hereinDescription.Bus interface 606 provides interface between bus 600 and receiver 601 and transmitter 603.Receiver 601 and transmitter603 can be the same element, i.e. transceiver, provide the unit for communicating over a transmission medium with various other devices.PlaceIt manages device 602 and is responsible for management bus 600 and common processing, and memory 604 can be used for storage processor 602 and execute behaviourUsed data when making.
Fourth aspect, based on the inventive concept with previous embodiment risk control method, the present invention also provides a kind of metersCalculation machine readable storage medium storing program for executing, is stored thereon with computer program, which realizes risk control described previously when being executed by processorThe step of either method processed method.
This specification is referring to the method, equipment (system) and computer program product according to this specification embodimentFlowchart and/or the block diagram describes.It should be understood that can be realized by computer program instructions every in flowchart and/or the block diagramThe combination of process and/or box in one process and/or box and flowchart and/or the block diagram.It can provide these computersProcessor of the program instruction to general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devicesTo generate a machine, so that generating use by the instruction that computer or the processor of other programmable data processing devices executeIn setting for the function that realization is specified in one or more flows of the flowchart and/or one or more blocks of the block diagramIt is standby.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spyDetermine 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 equipment, the commander equipment realize in one box of one or more flows of the flowchart and/or block diagram orThe 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 countingSeries of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer orThe instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram oneThe 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 basicProperty 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 asIt 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 artMind 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 technologiesWithin, then the present invention is also intended to include these modifications and variations.