CN109461068A - Judgment method, device, equipment and the computer readable storage medium of fraud - Google Patents
Judgment method, device, equipment and the computer readable storage medium of fraud Download PDFInfo
- Publication number
- CN109461068A CN109461068A CN201811066929.9A CN201811066929A CN109461068A CN 109461068 A CN109461068 A CN 109461068A CN 201811066929 A CN201811066929 A CN 201811066929A CN 109461068 A CN109461068 A CN 109461068A
- Authority
- CN
- China
- Prior art keywords
- fraud
- data
- model
- target
- type
- 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
- 238000000034 method Methods 0.000 title claims abstract description 46
- 230000006399 behavior Effects 0.000 claims abstract description 83
- 241000209202 Bromus secalinus Species 0.000 claims abstract description 37
- 238000012544 monitoring process Methods 0.000 claims abstract description 15
- 241001269238 Data Species 0.000 claims abstract description 10
- 230000008901 benefit Effects 0.000 claims description 17
- 238000004891 communication Methods 0.000 claims description 15
- 239000006185 dispersion Substances 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 4
- 238000010801 machine learning Methods 0.000 abstract 1
- 238000012512 characterization method Methods 0.000 description 13
- 230000008569 process Effects 0.000 description 9
- 238000010586 diagram Methods 0.000 description 5
- 238000001914 filtration Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 238000012216 screening Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000012954 risk control Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000004900 laundering Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The present invention discloses judgment method, device, equipment and the computer readable storage medium of a kind of fraud, the described method includes: reading the fraud data in default anti-fraud platform, and category filter is carried out to fraud data, generate various types of fraud sample datas；Each fraud sample data is respectively transmitted in default learning model, multiple fraud models are formed；When monitoring the operation behavior of user account, operation data is generated, and determines that target corresponding with operation data cheats model；Operation data is transferred in target fraud model, and cheats the similarity of target fraud data and operation data in model according to target, judges whether operation behavior is fraud.Because cheating fraud data of the sample data from default anti-fraud platform, make more acurrate to the judgement of fraud based on machine learning target fraud model generated；Artificial settings can be avoided, as fraud data update convenient for newly there is the judgement of fraud by cheating model simultaneously.
Description
Technical field
The invention mainly relates to financial institution's air control technical fields, specifically, being related to a kind of judgement side of fraud
Method, device, equipment and computer readable storage medium.
Background technique
Financial institution needs to carry out the risk control of the types such as money laundering, fraud, during operation to ensure finance
The interests of each user in mechanism.At present for fraud, usually judged by the related data being manually set, but because of reality
When fraud data involved in change greatly, cause setting related data there are larger differences with actual fraud data
It is different, and make to judge that there are errors；And the related data of setting needs artificial maintenance adjustment, makes promptly and accurately differentiate new appearance
Human cost is increased while fraud data.
Summary of the invention
The main object of the present invention is to provide the judgment method of fraud a kind of, device, equipment and computer-readable deposits
Storage media, it is intended to solve in the prior art by the judgement inaccuracy that data are manually set to fraud data, and increase people
The problem of power cost.
To achieve the above object, the present invention provides a kind of judgment method of fraud, the judgement side of the fraud
Method the following steps are included:
The fraud data in default anti-fraud platform are read, and category filter is carried out to the fraud data, are generated various
The fraud sample data of type；
Each fraud sample data is respectively transmitted in default learning model, multiple fraud models are formed；
When monitoring the operation behavior of user account, operation data is generated, and is determined in multiple fraud models
Target corresponding with the operation data cheats model；
The operation data is transferred in the target fraud model, and target in model is cheated according to the target and is taken advantage of
The similarity for cheating data and the operation data, judges whether the operation behavior is fraud.
Preferably, the similarity that target fraud data and the operation data in model are cheated according to the target,
Judge that the step of whether operation behavior is fraud includes:
The operation data and target fraud data are compared, determine that the operation data and the target are taken advantage of
Cheat the similarity numerical value of data；
Judge whether the similarity numerical value is greater than preset threshold, if the similarity numerical value is greater than the preset threshold,
Then determine the operation behavior for fraud.
Preferably, described to determine that the operation behavior includes: later for the step of fraud
User account corresponding with the operation behavior is transferred to default anti-fraud platform, for the default anti-fraud
Platform identifies the user account, generates authentication information and uploads；
When receiving the authentication information, the identification identifier in the authentication information is read, and according to the identification
Identifier determines whether the user account is adventure account；
If the user account is adventure account, the operation behavior is refused, and with the operation data and
Measures of dispersion between the target fraud data is updated target fraud model.
Preferably, described to determine target fraud model corresponding with the operation data in multiple fraud models
Step includes:
The operation data and preset keyword library are compared, determine multiple target keywords in the operation data,
And multiple target keywords are formed into the first keyword group；
Read the second keyword group corresponding with each fraud model, and by the first keyword group one by one with each institute
The matching of the second keyword group is stated, each matching angle value is generated；
Each matching angle value is compared, determines the maximum object matching angle value of numerical value, and will be with the target
It is determined as target fraud model with the fraud model that the corresponding second keyword group of angle value is belonged to.
Preferably, described that each fraud sample data is respectively transmitted in default learning model, form multiple frauds
The step of model includes:
The fraud type identification in each fraud sample data is read, and is judged whether there is and each fraud type
The corresponding type of mark cheats model；
The type corresponding with each fraud type identification cheats model if it exists, then according to each fraud sample
Notebook data carries out corresponding adjustment to each type fraud model；
The type corresponding with each fraud type identification cheats model if it does not exist, it is determined that each fraud
There is no the targets of the corresponding type fraud model to cheat type identification in type identification；
It determines the ownership fraud sample data in each target fraud type identification institute source, and each ownership is cheated
Sample data is transferred in default learning model, forms multiple fraud models.
Preferably, the step for carrying out corresponding adjustment to each type fraud model according to each fraud sample data
Suddenly include:
By each fraud sample data and corresponding each type fraud model comparison, each type fraud is determined
The variance data not having in model；
Each variance data is added in corresponding each type fraud model, model is cheated to each type
It is updated.
Preferably, include: before the step of fraud data read in default anti-fraud platform
The communication connection with default anti-fraud platform is established, when detection reaches preset time, to default anti-fraud platform
Send read requests.
In addition, to achieve the above object, the present invention also proposes a kind of judgment means of fraud, the fraud
Judgment means include:
Read module for reading the fraud data in default anti-fraud platform, and classifies to the fraud data
Screening, generates various types of fraud sample datas；
Transmission module forms multiple take advantage of for each fraud sample data to be respectively transmitted in default learning model
Cheat model；
Generation module generates operation data for when monitoring the operation behavior of user account, and described takes advantage of multiple
It cheats and determines that target corresponding with the operation data cheats model in model；
Judgment module for the operation data to be transferred in the target fraud model, and is taken advantage of according to the target
The similarity for cheating target fraud data and the operation data in model, judges whether the operation behavior is fraud.
In addition, to achieve the above object, the present invention also proposes a kind of judgement equipment of fraud, the fraud
Judge that equipment includes: memory, processor, communication bus and the judgement journey for the fraud being stored on the memory
Sequence；
The communication bus is for realizing the connection communication between processor and memory；
The processor is used to execute the determining program of the fraud, to perform the steps of
The fraud data in default anti-fraud platform are read, and category filter is carried out to the fraud data, are generated various
The fraud sample data of type；
Each fraud sample data is respectively transmitted in default learning model, multiple fraud models are formed；
When monitoring the operation behavior of user account, operation data is generated, and is determined in multiple fraud models
Target corresponding with the operation data cheats model；
The operation data is transferred in the target fraud model, and target in model is cheated according to the target and is taken advantage of
The similarity for cheating data and the operation data, judges whether the operation behavior is fraud.
In addition, to achieve the above object, it is described computer-readable the present invention also provides a kind of computer readable storage medium
Storage medium be stored with one perhaps more than one program the one or more programs can by one or one with
On processor execute to be used for:
The fraud data in default anti-fraud platform are read, and category filter is carried out to the fraud data, are generated various
The fraud sample data of type；
Each fraud sample data is respectively transmitted in default learning model, multiple fraud models are formed；
When monitoring the operation behavior of user account, operation data is generated, and is determined in multiple fraud models
Target corresponding with the operation data cheats model；
The operation data is transferred in the target fraud model, and target in model is cheated according to the target and is taken advantage of
The similarity for cheating data and the operation data, judges whether the operation behavior is fraud.
The judgment method of the fraud of the present embodiment, by the way that various types of fraud sample datas are transferred to default
It practises in model, generates the fraud model of each type；In the operation behavior for monitoring user account, operation data is generated, and
Determine that target corresponding with operation data cheats model in multiple fraud models；And then operation data is transferred to target fraud
In model, by target fraud model in cheat data and operation data similarity, judge user account operation behavior whether
For fraud.This programme is due to fraud sample data carries out category filter by the fraud data read from default fraud platform
Come, each fraud model generated is directly related to each history fraud, so that cheating model to fraud by target
Judgement it is more accurate；It is updated simultaneously because the fraud data in default fraud platform can be with the generation of fraud, so that
It is formed by fraud model also to automatically update, avoids artificials settings, be convenient for while reducing human cost to newly cheating
The accurate judgement of behavior.
Detailed description of the invention
Fig. 1 is the flow diagram of the judgment method first embodiment of fraud of the invention；
Fig. 2 is the functional block diagram of the judgment means first embodiment of fraud of the invention；
Fig. 3 is the device structure schematic diagram for the hardware running environment that present invention method is related to.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
The present invention provides a kind of judgment method of fraud.
Fig. 1 is please referred to, Fig. 1 is the flow diagram of the judgment method first embodiment of fraud of the present invention.In this reality
It applies in example, the judgment method of the fraud includes:
Step S10 reads the fraud data in default anti-fraud platform, and carries out category filter to the fraud data,
Generate various types of fraud sample datas；
The judgment method of fraud of the present invention is applied to the server of financial institution, is suitable for through server realization pair
The judgement of all kinds of frauds in financial institution.The risk control that fraud type is carried out in financial institution, is provided with default
Anti- fraud platform；All kinds of frauds are identified by all means by default anti-fraud platform, and to being identified as cheating
The fraud data of behavior are stored；As identification fraud in the identification card number of user, address name, morning in the late into the night it is different
Normal operating time, the abnormal operation behavior cancelled repeatedly etc..Book server and this default anti-fraud platform are established and are communicated to connect, with
Fraud data are read from default anti-fraud platform, and carry out subsequent fraud judgement according to the fraud data of reading.Cause
Fraud data in default anti-fraud platform are related to the fraud of identification, and the identification of fraud has not timing；
It is identified in certain section of time memory in fraud, and the fraud data in anti-fraud platform is updated；Another
There is no frauds in one period without being identified, without carrying out more to the fraud data in anti-fraud platform
Newly.To be previously provided with preset time, this is default for the fraud data in the more accurate default anti-fraud platform of reading
The duration of time is set in OK range by experiment；Avoid duration is too long from causing to cheat reading data not in time, Huo Zheshi
Length is too short to cause fraud reading data excessively frequently not yet to update.When detection reaches preset time, then to default anti-fraud
Platform sends read requests, reads fraud data therein with request.This read requests is received when presetting anti-fraud platform, then
The response to read requests is returned to server；If response reads the fraud number in default anti-fraud platform to allow to read
According to.Because default anti-fraud platform is for identifying various types of frauds, so that fraud data therein are related to
Multiple types may further include the normal data for storing non-fraud data as fraud data due to maloperation；To
When being read to fraud data, leading to the fraud data read includes multiple types and normal data.
It is more accurate in order to make to cheat data, after reading fraud data, category filter operation is carried out to fraud data；
Normal data is filtered out from fraud data first, then is classified to filtered fraud data, is generated various types of
Sample data is cheated, so that the corresponding a kind of fraud of a kind of fraud sample data.Wherein the filtering of normal data can be passed through
Map-reduce mode carries out, and map-reduce is that the task with mass data is divided into multiple small subtasks (map) simultaneously
After row executes, it is merged into the mode of result (reduce).Default anti-fraud platform is deposited to the fraud data of fraud
Chu Shi, according to the type of fraud be fraud data addition fraud type identification, such as respectively with the fraud of loan types,
All kinds of fraud data such as the fraud for insuring type, the fraud for mortgaging type, the fraud for assuring type are corresponding
Identifier F1, F2, F3, F4 etc..In filtering, all numbers with fraud type identification are filtered out from the fraud data of reading
It is to characterize the fraud data of all frauds according to, this all data, and it will be without the normal data mistake of fraud type identification
Filter.Hereafter in order to use the concrete type of each fraud data characterization fraud, to all data of filtering according to fraud type mark
Knowledge is classified；Data with identical fraud type identification are divided into same class, generates and characterizes each of various frauds
Class cheats data；Using this generate fraud data as fraud sample data, with according to cheat sample data to subsequent fraud
Behavior is judged.
Each fraud sample data is respectively transmitted in default learning model, forms multiple fraud moulds by step S20
Type；
Further, fraud is judged according to fraud sample data for the ease of subsequent, in the server of the present embodiment
It is previously provided with for trained default learning model.Default learning model can be supervised learning model, unsupervised learning
Model or semi-supervised learning model；Wherein supervised learning model first is marked data to form data set, then passes through number
Learn a kind of model out according to collection, goes to predict new data with this model learnt；Data do not need to mark in unsupervised learning model
Note, and gone automatically with algorithm model in discovery data some modes, such as cluster structure, hierarchical structure, sparse tree and figure
Etc., but accuracy is lower relative to supervised learning model；And in semi-supervised learning model, with unlabelled data and portion
Point flag data is come together training pattern, to improve the accuracy of model.After generating all kinds of fraud sample datas, taken advantage of all kinds of
Sample data is cheated as the data in default learning model for study；And this various types of fraud sample data is passed respectively
It is defeated to form multiple fraud models for judging various frauds into default learning model, and each fraud model is corresponding
A kind of fraud, to carry out sub-category accurate judgement to fraud.
Step S30 generates operation data, and in multiple fraud moulds when monitoring the operation behavior of user account
Determine that target corresponding with the operation data cheats model in type；
Further, the user in financial institution can (Application be applied by the website of financial institution, app
Program) etc. check operation to product information in financial institution or application process initiated to product；Such as to certain loan product
Information checked, and initiate application process, or application process etc. is initiated to certain insurance products.This generic operation needs user
Apply for user account in financial institution, operated by user account, and financial institution is provided with such operation behavior
Monitoring mechanism；In the such operation behavior for monitoring user account, by the corresponding data generation operations number of such operation behavior
According to.If the above-mentioned application to insurance products operates, corresponding operation data generated includes the user account of operation, applies being filled out
Name, ID card No., operated products type, operating time, number of operation modification for writing etc..Because of different fraud models pair
Applied to the judgement of different type fraud, thus after generating operation data, need to determine in multiple fraud models with
The corresponding fraud model of this operation data, cheats model using this fraud model as target, cheats model judgement behaviour by target
Whether the operation behavior for making data institute source belongs to fraud.Wherein, the determining and operation data pair in multiple fraud models
Answer target fraud model the step of include:
The operation data and preset keyword library are compared, determine multiple targets in the operation data by step S31
Keyword, and multiple target keywords are formed into the first keyword group；
Because each fraud model is related to the type of fraud, and the type of fraud passes through the key in fraud data
Word characterization；In view of the operation data for operating involved is numerous, need from wherein extracting effective keyword.The present embodiment is set
It is equipped with and is previously provided with preset keyword library, include the keyword of multiple characterization abnormal operation types in key word library, such as insure,
Loan, number of retries, operating time etc..Operation data and preset keyword library are compared, determination is existed simultaneously in operation data
With the keyword in preset keyword library, and using this keyword as in operation data target keywords formed the first keyword
Group characterizes abnormal operation type that may be present in operation data.
Step S32, reads the second keyword group corresponding with each fraud model, and by the first keyword group by
One and the matching of each second keyword group, generate each matching angle value；
Further, in order to characterize the applicable fraud type of each fraud model, in fraud model forming process
In, multiple keywords therein are read, the second keyword group is formed, is applicable in by this second keyword group come characterization model
Type；Keyword such as relevant to insurance insures the fraud model of type to characterize, and keyword relevant to mortgage loan carrys out table
Levy the fraud model of mortgage loan type.When determining target fraud model corresponding with operation data, each fraud mould is read
Second keyword group corresponding to type；And match the first keyword group of operation data and each second keyword group, it generates
The matching angle value of operation data and each fraud model characterizes the matching journey of operation data and each fraud model by each matching angle value
Degree.
Step S33 compares each matching angle value, determines the maximum object matching angle value of numerical value, and will be with institute
It states the fraud model that the corresponding second keyword group of object matching angle value is belonged to and is determined as target fraud model.
Further, after generating each matching angle value, each matching angle value is mutually compared, determines numerical value most
Big object matching angle value；Object matching angle value is generated by the first keyword group and the matching of a certain second keyword group, characterization behaviour
The matching degree highest for making data and the fraud model in this second keyword group institute source, with this fraud model to operation data pair
The operation behavior answered carries out fraudulent judgement, can make to judge more accurate.Specifically, generate object matching angle value second is crucial
Word group is the second keyword group corresponding with object matching angle value, and by this second keyword group corresponding with object matching angle value
The fraud model in institute source is determined as target fraud model.Because of the matching degree highest of target fraud model and operation data, mesh
Mark fraud model can reflect the characteristic of fraud data to the full extent；To be known using target fraud model to operation data
Do not judge, can make the behavior of operation data respective operations whether be fraud judgement accuracy it is higher.
The operation data is transferred in the target fraud model, and cheats model according to the target by step S40
The similarity of middle target fraud data and the operation data, judges whether the operation behavior is fraud.
Further, after the determining highest target fraud model with operation data matching degree, operation data is passed
It is defeated to be cheated in model to target, model is cheated by target and judges whether the operation behavior in operation data institute source is fraud row
For.Specifically, it cheats model because of target to be generated by the anti-fraud data study for cheating history fraud in platform, so that target
The target fraud data cheated in model reflect fraud；To in the judgment process, determine the mesh in target fraud model
Similarity between mark fraud data and operation data judges the corresponding operation behavior of operation data according to the size of similarity
It whether is fraud.If the similarity of the two is bigger, illustrate that the target in operation data and target fraud model is cheated
Data are closer, and the corresponding operation behavior of operation data more may be fraud；If similarity otherwise between the two is smaller,
Then illustrate that a possibility that operation behavior is fraud is lower.Specifically, according to target cheat model in target fraud data with
The similarity of operation data judges that the step of whether operation behavior is fraud includes:
The operation data and target fraud data are compared, determine the operation data and institute by step S41
State the similarity numerical value of target fraud data；
Understandably, the identical journey of the had data of both similarity characterizations between operation data and target fraud data
Degree determines identical data between the two and different data so that operation data and target fraud data be compared；And root
It is the similarity numerical value that can determine operation data and target fraud data according to identical data therein, as target fraud data include
5 data, and identical data is 3, then similarity numerical value is 3/5=0.6, has 60% data that may be in characterization operation data
Cheat data.
Step S42, judges whether the similarity numerical value is greater than preset threshold, if the similarity numerical value is greater than described pre-
If threshold value, then determine the operation behavior for fraud.
Further, in order to determine the size for cheating data included in operation data, it is previously provided with preset threshold；This is pre-
If threshold value is the numerical value by experiment test or experience setting, the amount of fraud data included in operation data is characterized higher than this number
Value then illustrates that fraud data more involved in operation data, the corresponding operation behavior of operation data are the possibility of fraud
Property is larger.After determining the similarity numerical value of target fraud data volume contained in characterization operation data, by this similarity numerical value
It is compared with preset threshold, judges whether similarity numerical value and preset threshold are greater than preset threshold；Illustrate to grasp if being not more than
Make amount containing target fraud data in data in the reasonable scope, and the operation behavior in this operation data institute source is determined as
Normal operation behavior；If more than then illustrating that the amount containing target fraud data in operation data is larger, this operation data is corresponding
Operation behavior there may be risks, and in order to avoid risk, the operation behavior in this operation data institute source is judged to cheating
Behavior is to carry out air control.
The judgment method of the fraud of the present embodiment, by the way that various types of fraud sample datas are transferred to default
It practises in model, generates the fraud model of each type；In the operation behavior for monitoring user account, operation data is generated, and
Determine that target corresponding with operation data cheats model in multiple fraud models；And then operation data is transferred to target fraud
In model, by target fraud model in cheat data and operation data similarity, judge user account operation behavior whether
For fraud.This programme is due to fraud sample data carries out category filter by the fraud data read from default fraud platform
Come, each fraud model generated is directly related to each history fraud, so that cheating model to fraud by target
Judgement it is more accurate；It is updated simultaneously because the fraud data in default fraud platform can be with the generation of fraud, so that
It is formed by fraud model also to automatically update, avoids artificials settings, be convenient for while reducing human cost to newly cheating
The accurate judgement of behavior.
Further, described to determine the operation behavior in another embodiment of judgment method of fraud of the present invention
Include: later for the step of fraud
User account corresponding with the operation behavior is transferred to default anti-fraud platform, for described pre- by step S43
If anti-fraud platform identifies the user account, generates authentication information and upload；
Understandably, after cheating model by target and judging operation behavior for fraud, in order to further determine
Whether operation behavior is fraud, and the present embodiment is provided with the mechanism that default anti-fraud platform carries out secondary-confirmation.Tool
Body, will corresponding with operation behavior user account, i.e., be transferred to default counter take advantage of in the user account that financial institution is operated
It cheats in platform, this user account is identified in default anti-fraud platform.Identify Shi Xianxiang financial institution request user
The registration informations such as identity information, name, telephone number that account is filled in registration process by registration information and preset counter take advantage of
Blacklist in swindleness platform compares, and judges whether have any one information to be present in blacklist in registration information；Simultaneously also
Artificial evidence obtaining identification is combined for registration information, generates authentication information.Specifically, when identity information is present in blacklist,
Because identity information is the information for characterizing user's uniqueness, and without identification of artificially collecting evidence, directly generating user account is risk
The authentication information of account；When name or telephone number are present in blacklist, because name or telephone number be easy to appear repetition or
The phenomenon that usurping, and artificial evidence obtaining identification is carried out, authentication information is generated according to the result of artificial evidence obtaining identification, to ensure to identify letter
The accuracy of breath.And when being present in blacklist in registration information without any one information, then illustrate the corresponding use of registration information
Family did not carry out risk operations before this；It is same to carry out artificial evidence obtaining identification in order to ensure safety, and according to artificial evidence obtaining identification
Result generate authentication information.By this generate authentication information upload onto the server, in the server combining target fraud model,
Default anti-fraud platform and artificial evidence obtaining three determine operation behavior, it can be ensured that the accuracy of determined result.
Step S44 reads the identification identifier in the authentication information when receiving the authentication information, and according to
The identification identifier determines whether the user account is adventure account；
Further, because authentication information includes that user account is adventure account and is not two class of adventure account, thus
When generating authentication information, in order to distinguish to the two, two category informations are distributed with different identification identifiers.It is identified when receiving
When information, read the identification identifier that carries in authentication information, this identify identifier for characterization user account be adventure account also
It is the identifier of non-adventure account；To identify identifier according to this, determine whether user is adventure account.
Step S45 refuses the operation behavior if the user account is adventure account, and with the behaviour
The measures of dispersion made between data and target fraud data is updated target fraud model.
It is non-adventure account when identifying identifier as characterization user account, i.e., through default anti-fraud platform and artificial identification
When risk is not present in user account, but because being fraud through target fraud model decision behavior；So as to be grasped
The user account of work is not present risk before this, but this time there may be risks for operation behavior, and to this operation behavior into
Row key monitoring, to prevent from cheating.And when identify that identifier be user account is adventure account, then explanation is cheating mould through target
While type judges operation behavior for fraud, then identify through default anti-fraud platform and artificially the user's account operated
There are risks at family, to refuse operation behavior, cause to cheat to avoid operation behavior.Simultaneously therefore operation behavior behaviour
The similarity numerical value made between the target fraud data in data and target fraud model is greater than preset threshold, endless between the two
It is exactly the same；Using this not exactly the same data as measures of dispersion between the two, target fraud model is carried out with this measures of dispersion
Automatically update, with expand target fraud model identification types, make target fraud model it is subsequent can more accurate judgement take advantage of
Swindleness behavior.
Further, described by each fraud sample in another embodiment of judgment method of fraud of the present invention
The step of data are respectively transmitted in default learning model, form multiple fraud models include:
Step S21, read it is each it is described fraud sample data in fraud type identification, and judge whether there is with it is each described
It cheats the corresponding type of type identification and cheats model；
Understandably, due to the fraud data in default anti-fraud platform can be with the update of identified fraud more
Newly, and the fraud that is identified by fraud model other than being determined, also relates to the fraud identified by other approach.
And the update fraud for being identified by other approach, because not being related to cheating model, and this part is made to update fraud
The fraud data of behavior are not present in fraud model, can not be automatically updated as the measures of dispersion of fraud model.It needs
Fraud model is generated for the fraud data of this type, or is added in original fraud model, in order to subsequently through fraud
Model identifies such fraud.Specifically, when reaching preset time, the fraud in default anti-fraud platform is read
Data, this fraud data are the new fraud data after a upper preset time is read, and are carried out to this new fraud data
Filtering screening generates fraud sample data.Because being provided with the fraud type mark of characterization fraud type in fraud sample data
Know, and cheat model and generated by the training of fraud sample data, so that also being carried in fraud model corresponding with fraud sample data
Fraud type identification.After generating various types of fraud sample datas, the fraud class in this each fraud sample data is read
Type mark, and each fraud type identification that this is read is compared with the fraud type identification carried in fraud model one by one, judgement
It whether there is fraud type corresponding with each fraud type identification of reading in the fraud type identification that fraud model carries
Mark, that is, judge whether the fraud type of each fraud sample data characterization has corresponding fraud model.This is corresponded to
Fraud model as type cheat model, with characterize carry reading fraud type identification fraud type, with and other
Fraud model is distinguished.
Step S22, the type corresponding with each fraud type identification cheats model if it exists, then according to each institute
It states fraud sample data and corresponding adjustment is carried out to each type fraud model；
Further, when each fraud type mark for judging to have with reading in the fraud type identification for cheating model carrying
When knowing corresponding fraud type identification, then illustrates the fraud type that each fraud sample data is characterized, exist and correspond to
Type fraud model this all kinds of fraud is judged.Therefore fraud sample data is in default anti-fraud platform pre-
If the new fraud sample data that is stored in the time, thus need using this each fraud sample data to all types of fraud models into
Row adjustment, and be adjusted according to fraud sample data and the corresponding relationship of type fraud model.Type such as is insured for characterization
The fraud sample data a of fraud, and model A is cheated for the type for judging Insurance Fraud behavior, the two has identical
Fraud type identification, i.e., both there is corresponding relationship, to adjust using fraud sample data a to type fraud model
It is whole.Specifically, carrying out the step of correspondence adjusts to all types of fraud models according to each fraud sample data includes:
Each fraud sample data and corresponding each type fraud model comparison are determined each institute by step S221
State the variance data not having in type fraud model；
Specifically, in the corresponding relationship according to fraud sample data and type fraud model, type fraud model is carried out
During adjustment；According to corresponding relationship between the two, fraud sample data and corresponding type fraud model are carried out pair
Than determination is present in fraud sample data, and the variance data in type fraud model may be not present.This variance data is default
The new fraud sample data that anti-fraud platform stores within a preset time, characterizes emerging fraud.
Each variance data is added in corresponding each type fraud model, to each class by step S222
Type fraud model is updated.
Further, this variance data for characterizing emerging fraud is added to corresponding type and cheats model
In, type fraud model is updated, so that the wider model of fraud type that type fraud model can judge, subsequent right
The judgement of fraud is more acurrate.
Step S23, the type corresponding with each fraud type identification cheats model if it does not exist, it is determined that each
There is no the targets of the corresponding type fraud model to cheat type identification in the fraud type identification；
And if judging that there is no equal with each fraud type identification of reading in the fraud type identification of fraud model carrying
Corresponding fraud type identification then illustrates the fraud type that each fraud sample data is characterized, corresponding class is not present
Type fraud model judges this all kinds of fraud；Sample data is cheated for part, there is no types to cheat model
The fraud characterized to it judges.And it needs to generate new fraud model for this part fraud sample data.It will
During each fraud type identification read is compared with the fraud type identification carried in fraud model one by one, it is not present in taking advantage of
Each fraud type identification for cheating the entrained fraud type identification of model is determined as multiple target fraud type identifications, to characterize not
Present pattern is cheated model and is judged the fraud of this multiple target frauds type identification institute source fraud sample data.
Step S24 determines the ownership fraud sample data in each target fraud type identification institute source, and will be each described
Ownership fraud sample data is transferred in default learning model, forms multiple fraud models.
Further, all kinds of fraud sample datas that will carry this all kinds of target fraud type identification, are determined as each mesh
The ownership in mark fraud type identification institute source cheats sample data；This all kinds of ownership is cheated into sample data as default
The data in model for study are practised, is transferred in default learning model and is trained, are formed for judging a variety of new fraud rows
For multiple fraud models, to carry out classification judgement to various frauds subsequent.
In addition, referring to figure 2., the present invention provides a kind of judgment means of fraud, in sentencing for fraud of the present invention
In disconnected device first embodiment, the judgment means of the fraud include:
Read module 10 for reading the fraud data in default anti-fraud platform, and divides the fraud data
Class screening, generates various types of fraud sample datas；
Transmission module 20 is formed multiple for each fraud sample data to be respectively transmitted in default learning model
Cheat model；
Generation module 30, for generating operation data, and multiple described when monitoring the operation behavior of user account
It cheats and determines that target corresponding with the operation data cheats model in model；
Judgment module 40, for the operation data to be transferred in the target fraud model, and according to the target
The similarity for cheating target fraud data and the operation data in model, judges whether the operation behavior is fraud.
The judgment means of the fraud of the present embodiment, by read module 10 and transmission module 20 by the various types of of reading
Type fraud sample data is transferred in default learning model, generates the fraud model of each type；Generation module 30 is monitoring
When the operation behavior of user account, operation data is generated, and determines target corresponding with operation data in multiple fraud models
Cheat model；And then operation data is transferred in target fraud model, judgment module 40 is by cheating number in target fraud model
According to the similarity with operation data, judge whether the operation behavior of user account is fraud.This programme is because cheating sample number
According to by from default fraud platform the fraud data that read carry out from category filter, each fraud model generated is directly and respectively
History fraud is related, so that it is more accurate to the judgement of fraud to cheat model by target；Simultaneously because of default fraud
Fraud data in platform can be updated with the generation of fraud, also automatically updated, kept away so that being formed by fraud model
Exempt to be manually set, convenient for newly there is the accurate judgement of fraud while reducing human cost.
Further, in another embodiment of judgment means of fraud of the present invention, the judgment module includes:
Comparison unit determines the operand for comparing the operation data and target fraud data
According to the similarity numerical value with target fraud data；
Judging unit, for judging whether the similarity numerical value is greater than preset threshold, if the similarity numerical value is greater than
The preset threshold then determines the operation behavior for fraud.
Further, in another embodiment of judgment means of fraud of the present invention, the judging unit is also used to:
User account corresponding with the operation behavior is transferred to default anti-fraud platform, for the default anti-fraud
Platform identifies the user account, generates authentication information and uploads；
When receiving the authentication information, the identification identifier in the authentication information is read, and according to the identification
Identifier determines whether the user account is adventure account；
If the user account is adventure account, the operation behavior is refused, and with the operation data and
Measures of dispersion between the target fraud data is updated target fraud model.
Further, in another embodiment of judgment means of fraud of the present invention, the generation module includes:
Unit is formed, for comparing the operation data and preset keyword library, is determined more in the operation data
A target keywords, and multiple target keywords are formed into the first keyword group；
First reading unit, for reading the second keyword group corresponding with each fraud model, and by described first
Keyword group is matched with each second keyword group one by one, generates each matching angle value；
First determination unit determines the maximum object matching angle value of numerical value for comparing each matching angle value,
And the fraud model that the second keyword group corresponding with the object matching angle value is belonged to is determined as target fraud mould
Type.
Further, in another embodiment of judgment means of fraud of the present invention, the transmission module includes:
Second reading unit for reading the fraud type identification in each fraud sample data, and judges whether to deposit
Model is cheated in type corresponding with each fraud type identification；
Adjustment unit cheats model for the type corresponding with each fraud type identification if it exists, then root
Corresponding adjustment is carried out to each type fraud model according to each fraud sample data；
Second determination unit cheats mould for the type corresponding with each fraud type identification if it does not exist
Type, it is determined that there is no the targets of the corresponding type fraud model to cheat type identification in each fraud type identification；
Transmission unit, for determining that the ownership in each target fraud type identification institute source cheats sample data, and will
Each ownership fraud sample data is transferred in default learning model, forms multiple fraud models.
Further, in another embodiment of judgment means of fraud of the present invention, the adjustment unit is also used to:
By each fraud sample data and corresponding each type fraud model comparison, each type fraud is determined
The variance data not having in model；
Each variance data is added in corresponding each type fraud model, model is cheated to each type
It is updated.
Further, in another embodiment of judgment means of fraud of the present invention, the judgement of the fraud is filled
It sets further include:
Sending module, for establishing the communication connection with default anti-fraud platform, when detection reaches preset time, to pre-
If anti-fraud platform sends read requests.
Wherein, each virtual functions module of the judgment means of above-mentioned fraud is stored in sentencing for fraud shown in Fig. 3
In the memory 1005 of disconnected equipment, when processor 1001 executes the determining program of fraud, realize each in embodiment illustrated in fig. 2
The function of a module.
Referring to Fig. 3, Fig. 3 is the device structure schematic diagram for the hardware running environment that present invention method is related to.
The judgement equipment of fraud of the embodiment of the present invention can be PC (personal computer, personal computer),
It is also possible to the terminal devices such as smart phone, tablet computer, E-book reader, portable computer.
As shown in figure 3, the judgement equipment of the fraud may include: processor 1001, such as CPU (Central
Processing Unit, central processing unit), memory 1005, communication bus 1002.Wherein, communication bus 1002 for realizing
Connection communication between processor 1001 and memory 1005.Memory 1005 can be high-speed RAM (random access
Memory, random access memory), it is also possible to stable memory (non-volatile memory), such as disk storage
Device.Memory 1005 optionally can also be the storage device independently of aforementioned processor 1001.
Optionally, the judgement equipment of the fraud can also include user interface, network interface, camera, RF
(Radio Frequency, radio frequency) circuit, sensor, voicefrequency circuit, WiFi (Wireless Fidelity, WiMAX) mould
Block etc..User interface may include display screen (Display), input unit such as keyboard (Keyboard), and optional user connects
Mouth can also include standard wireline interface and wireless interface.Network interface optionally may include the wireline interface, wireless of standard
Interface (such as WI-FI interface).
It will be understood by those skilled in the art that the judgement device structure of fraud shown in Fig. 3 is not constituted to taking advantage of
The restriction of the judgement equipment of swindleness behavior may include perhaps combining certain components or not than illustrating more or fewer components
Same component layout.
As shown in figure 3, as may include operating system, net in a kind of memory 1005 of computer readable storage medium
Network communication module and the determining program of fraud.Operating system be manage and control fraud judgement device hardware and
The program of software resource supports the operation of the determining program and other softwares and/or program of fraud.Network communication module
For realizing the communication between each component in the inside of memory 1005, and with other hardware in the judgement equipment of fraud and soft
It is communicated between part.
In the judgement equipment of fraud shown in Fig. 3, processor 1001 is used to execute to store in memory 1005
The determining program of fraud realizes the step in each embodiment of the judgment method of above-mentioned fraud.
The present invention provides a kind of computer readable storage medium, the computer-readable recording medium storage have one or
More than one program of person, the one or more programs can also be executed by one or more than one processor with
Step in each embodiment of judgment method for realizing above-mentioned fraud.
It should also be noted that, herein, the terms "include", "comprise" or its any other variant are intended to non-
It is exclusive to include, so that the process, method, article or the device that include a series of elements not only include those elements,
It but also including other elements that are not explicitly listed, or further include solid by this process, method, article or device
Some elements.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including
There is also other identical elements in the process, method of the element, article or device.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words does the prior art
The part contributed out can be embodied in the form of software products, which is stored in one as described above
In computer readable storage medium (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that a terminal device (can
To be mobile phone, computer, server or the network equipment etc.) execute method described in each embodiment of the present invention.
The above description is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all at this
Under the design of invention, using equivalent structure transformation made by description of the invention and accompanying drawing content, or directly/it is used in it indirectly
He is included in scope of patent protection of the invention relevant technical field.
Claims (10)
1. a kind of judgment method of fraud, which is characterized in that the judgment method of the fraud the following steps are included:
The fraud data in default anti-fraud platform are read, and category filter is carried out to the fraud data, generate various types
Fraud sample data；
Each fraud sample data is respectively transmitted in default learning model, multiple fraud models are formed；
When monitoring the operation behavior of user account, operation data is generated, and the determining and institute in multiple fraud models
State the corresponding target fraud model of operation data；
The operation data is transferred in the target fraud model, and target in model is cheated according to the target and cheats number
According to the similarity with the operation data, judge whether the operation behavior is fraud.
2. the judgment method of fraud as described in claim 1, which is characterized in that described to cheat model according to the target
The similarity of middle target fraud data and the operation data, judges the step of whether operation behavior is fraud packet
It includes:
The operation data and target fraud data are compared, determine the operation data and target fraud number
According to similarity numerical value；
Judge whether the similarity numerical value is greater than preset threshold, if the similarity numerical value is greater than the preset threshold, sentences
The fixed operation behavior is fraud.
3. the judgment method of fraud as claimed in claim 2, which is characterized in that described to determine that the operation behavior is to take advantage of
Include: after the step of swindleness behavior
User account corresponding with the operation behavior is transferred to default anti-fraud platform, for the default anti-fraud platform
The user account is identified, authentication information is generated and is uploaded；
When receiving the authentication information, the identification identifier in the authentication information is read, and identify according to the identification
Symbol, determines whether the user account is adventure account；
If the user account be adventure account, the operation behavior is refused, and with the operation data with it is described
Measures of dispersion between target fraud data is updated target fraud model.
4. the judgment method of fraud as described in claim 1, which is characterized in that described in multiple fraud models
Determine that the step of target corresponding with the operation data cheats model includes:
The operation data and preset keyword library are compared, determine multiple target keywords in the operation data, and will
Multiple target keywords form the first keyword group；
Read the second keyword group corresponding with each fraud model, and by the first keyword group one by one with each described the
The matching of two keyword groups, generates each matching angle value；
Each matching angle value is compared, determines the maximum object matching angle value of numerical value, and will be with the object matching degree
It is worth the fraud model that the corresponding second keyword group is belonged to and is determined as target fraud model.
5. the judgment method of fraud as described in claim 1, which is characterized in that described by each fraud sample data
The step of being respectively transmitted in default learning model, forming multiple fraud models include:
The fraud type identification in each fraud sample data is read, and is judged whether there is and each fraud type identification
Corresponding type cheats model；
The type corresponding with each fraud type identification cheats model if it exists, then according to each fraud sample number
Corresponding adjustment is carried out according to each type fraud model；
The type corresponding with each fraud type identification cheats model if it does not exist, it is determined that each fraud type
There is no the targets of the corresponding type fraud model to cheat type identification in mark；
It determines the ownership fraud sample data in each target fraud type identification institute source, and each ownership is cheated into sample
Data are transferred in default learning model, form multiple fraud models.
6. the judgment method of fraud as claimed in claim 5, which is characterized in that described according to each fraud sample number
Include: according to the step of carrying out corresponding adjustment to each type fraud model
By each fraud sample data and corresponding each type fraud model comparison, each type fraud model is determined
In the variance data that does not have；
Each variance data is added in corresponding each type fraud model, each type fraud model is carried out
It updates.
7. the judgment method of fraud as claimed in any one of claims 1 to 6, which is characterized in that described read presets counter take advantage of
Include: before the step of cheating the fraud data in platform
The communication connection with default anti-fraud platform is established, when detection reaches preset time, is sent to default anti-fraud platform
Read requests.
8. a kind of judgment means of fraud, which is characterized in that the judgment means of the fraud include:
Read module for reading the fraud data in default anti-fraud platform, and carries out category filter to the fraud data,
Generate various types of fraud sample datas；
Transmission module forms multiple fraud moulds for each fraud sample data to be respectively transmitted in default learning model
Type；
Generation module, for generating operation data, and in multiple fraud moulds when monitoring the operation behavior of user account
Determine that target corresponding with the operation data cheats model in type；
Judgment module for the operation data to be transferred in the target fraud model, and cheats mould according to the target
The similarity of target fraud data and the operation data, judges whether the operation behavior is fraud in type.
9. a kind of judgement equipment of fraud, which is characterized in that the judgement equipment of the fraud includes: memory, place
Manage device, communication bus and the determining program for the fraud being stored on the memory；
The communication bus is for realizing the connection communication between processor and memory；
The processor is used to execute the determining program of the fraud, to realize as described in any one of claim 1-7
Fraud judgment method the step of.
10. a kind of computer readable storage medium, which is characterized in that be stored with fraud row on the computer readable storage medium
For determining program, when the determining program of the fraud is executed by processor realize such as any one of claim 1-7 institute
The step of judgment method for the fraud stated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811066929.9A CN109461068A (en) | 2018-09-13 | 2018-09-13 | Judgment method, device, equipment and the computer readable storage medium of fraud |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811066929.9A CN109461068A (en) | 2018-09-13 | 2018-09-13 | Judgment method, device, equipment and the computer readable storage medium of fraud |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109461068A true CN109461068A (en) | 2019-03-12 |
Family
ID=65606754
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811066929.9A Pending CN109461068A (en) | 2018-09-13 | 2018-09-13 | Judgment method, device, equipment and the computer readable storage medium of fraud |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109461068A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110033170A (en) * | 2019-03-14 | 2019-07-19 | 阿里巴巴集团控股有限公司 | Identify the method and device of risk businessman |
| CN110569906A (en) * | 2019-09-10 | 2019-12-13 | 京东数字科技控股有限公司 | Data processing method, data processing apparatus, and computer-readable storage medium |
| CN111125658A (en) * | 2019-12-31 | 2020-05-08 | 深圳市分期乐网络科技有限公司 | Method, device, server and storage medium for identifying fraudulent users |
| WO2020253116A1 (en) * | 2019-06-20 | 2020-12-24 | 深圳壹账通智能科技有限公司 | Batch data execution method, device, storage medium, and member host in cluster |
| CN113596844A (en) * | 2021-07-29 | 2021-11-02 | 恒安嘉新（北京）科技股份公司 | Early warning method, device, medium and electronic equipment based on data information |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567788A (en) * | 2010-12-28 | 2012-07-11 | 中国移动通信集团重庆有限公司 | Real-time identification system and real-time identification method for fraudulent practice in communication services |
| CN106447333A (en) * | 2016-11-29 | 2017-02-22 | 中国银联股份有限公司 | Fraudulent trading detection method and server |
| CN106506454A (en) * | 2016-10-10 | 2017-03-15 | 江苏通付盾科技有限公司 | Fraud business recognition method and device |
| CN106600413A (en) * | 2015-10-19 | 2017-04-26 | 阿里巴巴集团控股有限公司 | Cheat recognition method and system |
| CN107240024A (en) * | 2017-05-22 | 2017-10-10 | 中国平安人寿保险股份有限公司 | The anti-fraud recognition methods of settlement of insurance claim and device |
| US20180253755A1 (en) * | 2016-05-24 | 2018-09-06 | Tencent Technology (Shenzhen) Company Limited | Method and apparatus for identification of fraudulent click activity |
-
2018
- 2018-09-13 CN CN201811066929.9A patent/CN109461068A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567788A (en) * | 2010-12-28 | 2012-07-11 | 中国移动通信集团重庆有限公司 | Real-time identification system and real-time identification method for fraudulent practice in communication services |
| CN106600413A (en) * | 2015-10-19 | 2017-04-26 | 阿里巴巴集团控股有限公司 | Cheat recognition method and system |
| US20180253755A1 (en) * | 2016-05-24 | 2018-09-06 | Tencent Technology (Shenzhen) Company Limited | Method and apparatus for identification of fraudulent click activity |
| CN106506454A (en) * | 2016-10-10 | 2017-03-15 | 江苏通付盾科技有限公司 | Fraud business recognition method and device |
| CN106447333A (en) * | 2016-11-29 | 2017-02-22 | 中国银联股份有限公司 | Fraudulent trading detection method and server |
| CN107240024A (en) * | 2017-05-22 | 2017-10-10 | 中国平安人寿保险股份有限公司 | The anti-fraud recognition methods of settlement of insurance claim and device |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110033170A (en) * | 2019-03-14 | 2019-07-19 | 阿里巴巴集团控股有限公司 | Identify the method and device of risk businessman |
| WO2020181910A1 (en) * | 2019-03-14 | 2020-09-17 | 阿里巴巴集团控股有限公司 | Method and device for identifying a risk merchant |
| TWI724552B (en) * | 2019-03-14 | 2021-04-11 | 開曼群島商創新先進技術有限公司 | Method and device for identifying risky merchants |
| US11379845B2 (en) | 2019-03-14 | 2022-07-05 | Advanced New Technologies Co., Ltd. | Method and device for identifying a risk merchant |
| WO2020253116A1 (en) * | 2019-06-20 | 2020-12-24 | 深圳壹账通智能科技有限公司 | Batch data execution method, device, storage medium, and member host in cluster |
| CN110569906A (en) * | 2019-09-10 | 2019-12-13 | 京东数字科技控股有限公司 | Data processing method, data processing apparatus, and computer-readable storage medium |
| CN110569906B (en) * | 2019-09-10 | 2022-08-09 | 京东科技控股股份有限公司 | Data processing method, data processing apparatus, and computer-readable storage medium |
| CN111125658A (en) * | 2019-12-31 | 2020-05-08 | 深圳市分期乐网络科技有限公司 | Method, device, server and storage medium for identifying fraudulent users |
| CN111125658B (en) * | 2019-12-31 | 2024-03-22 | 深圳市分期乐网络科技有限公司 | Method, apparatus, server and storage medium for identifying fraudulent user |
| CN113596844A (en) * | 2021-07-29 | 2021-11-02 | 恒安嘉新（北京）科技股份公司 | Early warning method, device, medium and electronic equipment based on data information |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109461068A (en) | Judgment method, device, equipment and the computer readable storage medium of fraud | |
| CN110197315B (en) | Risk assessment method, apparatus and storage medium thereof | |
| CN108665355A (en) | Financial product recommends method, apparatus, equipment and computer storage media | |
| CN106104555B (en) | For protecting the behavioural analysis of peripheral equipment | |
| US20210019657A1 (en) | Fraud detection and risk assessment method, system, device, and storage medium | |
| US12118552B2 (en) | User profiling based on transaction data associated with a user | |
| CN104836781B (en) | Distinguish the method and device for accessing user identity | |
| CN110442712B (en) | Risk determination method, risk determination device, server and text examination system | |
| CN110689438A (en) | Enterprise financial risk scoring method and device, computer equipment and storage medium | |
| CN108399565A (en) | Financial product recommendation apparatus, method and computer readable storage medium | |
| CN109978033A (en) | The method and apparatus of the building of biconditional operation people's identification model and biconditional operation people identification | |
| CN111522724B (en) | Method and device for determining abnormal account number, server and storage medium | |
| CN109600336A (en) | Store equipment, identifying code application method and device | |
| CN109684364A (en) | The problem of being drawn a portrait based on user processing method, device, equipment and storage medium | |
| CN109194689A (en) | Abnormal behaviour recognition methods, device, server and storage medium | |
| KR20190002593A (en) | Account fulfillment method and server | |
| CN112634017A (en) | Remote card opening activation method and device, electronic equipment and computer storage medium | |
| CN110457601A (en) | The recognition methods and device of social account, storage medium and electronic device | |
| CN110675252A (en) | Risk assessment method and device, electronic equipment and storage medium | |
| TW201923678A (en) | Shared product recommendation method and apparatus, and electronic device | |
| CN110399818A (en) | A kind of method and apparatus of risk profile | |
| CN111784360B (en) | Anti-fraud prediction method and system based on network link backtracking | |
| CN110197375A (en) | A kind of similar users recognition methods, device, similar users identification equipment and medium | |
| CN107832925A (en) | Internet content risk evaluating method, device and server | |
| CN110428342B (en) | Data restoration method, server, customer service side and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 40000890 Country of ref document: HK |
|
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190312 |