CN106503562A - A kind of Risk Identification Method and device - Google Patents
A kind of Risk Identification Method and device Download PDFInfo
- Publication number
- CN106503562A CN106503562A CN201510560860.5A CN201510560860A CN106503562A CN 106503562 A CN106503562 A CN 106503562A CN 201510560860 A CN201510560860 A CN 201510560860A CN 106503562 A CN106503562 A CN 106503562A
- Authority
- CN
- China
- Prior art keywords
- account
- value
- sample
- parameter
- risk assessment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/50—Monitoring users, programs or devices to maintain the integrity of platforms, e.g. of processors, firmware or operating systems
- G06F21/57—Certifying or maintaining trusted computer platforms, e.g. secure boots or power-downs, version controls, system software checks, secure updates or assessing vulnerabilities
- G06F21/577—Assessing vulnerabilities and evaluating computer system security
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/285—Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system
Abstract
The application is related to technical field of data processing, more particularly to a kind of Risk Identification Method and device, relatively low in order to solve the problems, such as the accuracy for carrying out risk identification in prior art.The Risk Identification Method that the embodiment of the present application is provided includes：The Business Processing for account to be identified that server receive user terminal sends is asked；The first machine sort model that the characteristic value input of multiple fisrt feature parameters of the account to be identified is trained, obtains the first risk assessment value；The fisrt feature parameter is the parameter of the corresponding non-accumulative behavioural characteristic of the reflection account；By the first risk assessment value, and the second machine sort model that the characteristic value input of at least one second feature parameter of the account to be identified is trained, obtain the second risk assessment value；The second feature parameter is the parameter of the corresponding accumulative behavioural characteristic of the reflection account；Based on the second risk assessment value for obtaining, risk identification is carried out to Business Processing request.
Description
Technical field
The application is related to technical field of data processing, more particularly to a kind of Risk Identification Method and device.
Background technology
In risk identification, server generally needs to train machine sort model based on history process record,
The machine sort model that the characteristic value input of the various features parameter of account to be identified is trained, based on model
The result of output carries out risk identification.These characteristic parameters have the parameter of the accumulative behavioural characteristic of reflection account,
The Internet protocol (Internet Protocol, IP) of initiating business request is changed in such as one month account
Number of times of address etc., also reflects the parameter of non-accumulative behavioural characteristic, such as Account Type etc..
At present, typically by the be possible to characteristic parameter for affecting risk identification result all directly as machine point
The Consideration of class model.But, reflect that the parameter of non-accumulative behavioural characteristic generally occurs noise data,
That is abnormal data, these noise datas can produce impact to model accuracy and stability, so as to have a strong impact on wind
The accuracy of dangerous recognition result.
In order to reduce the impact of noise data, by non-for this reflection accumulative behavioural characteristic in some processing modes
Parameter carries out screening out process, but this mode often omits a lot of useful informations, and then causes to carry out risk
The accuracy of identification is relatively low.
Content of the invention
The embodiment of the present application provides a kind of Risk Identification Method and device, in order to solve to enter sector-style in prior art
The relatively low problem of the dangerous accuracy for recognizing.
The embodiment of the present application provides a kind of Risk Identification Method, including：
The Business Processing for account to be identified that server receive user terminal sends is asked；
The first machine that the characteristic value input of multiple fisrt feature parameters of the account to be identified is trained
Disaggregated model, obtains the first risk assessment value；The fisrt feature parameter is corresponding non-for the reflection account
The parameter of accumulative behavioural characteristic；
By the first risk assessment value, and the spy of at least one second feature parameter of the account to be identified
The second machine sort model that value indicative input is trained, obtains the second risk assessment value；The second feature ginseng
Parameter of the number for the corresponding accumulative behavioural characteristic of the reflection account；
Based on the second risk assessment value for obtaining, risk identification is carried out to Business Processing request.
Alternatively, the first machine sort model is trained according to following steps：
Obtain the characteristic parameter information of the sample account of multiple known Account Types；Wherein, each sample account
Characteristic parameter information include the characteristic value of multiple fisrt feature parameters；The Account Type includes secured account
And adventure account；
Based on the characteristic value of multiple fisrt feature parameters of each sample account in the plurality of sample account, with
And the Account Type of each sample account, train the first machine sort model.
Alternatively, the spy based on multiple fisrt feature parameters of each sample account in the plurality of sample account
Value indicative, and the Account Type of each sample account, train the first machine sort model, including：
For each the sample account in the plurality of sample account, with multiple fisrt feature of the sample account
Input value of the characteristic value of parameter as the first machine sort model, with the Account Type of the sample account
Corresponding risk assessment value trains described first as the target output value of the first machine sort model
Machine sort model.
Alternatively, the characteristic parameter information of each sample account also includes the spy of at least one second feature parameter
Value indicative；The second machine sort model is trained according to following steps：
First machine that the characteristic value input of the various fisrt feature parameters of each sample account is trained
Device disaggregated model, obtains the first risk assessment value；
Based in the plurality of sample account, each sample account distinguishes corresponding various second feature parameters
Characteristic value, the first risk assessment value and Account Type, train the second machine sort model.
Alternatively, in based on the plurality of sample account, corresponding various second spies of each sample account difference
Characteristic value, the first risk assessment value and the Account Type of parameter is levied, second machine is trained
Disaggregated model, including：
For each the sample account in the plurality of sample account, with the various second feature of the sample account
The input value of the characteristic value of parameter and the first risk assessment value as the second machine sort model, with the sample
The corresponding risk assessment value of the Account Type of this account is exported as the target of the second machine sort model
Value, trains the second machine sort model.
The embodiment of the present application provides a kind of risk identification device, including：
Receiver module, asks for the Business Processing for account to be identified that receive user terminal sends；
First processing module, is input into for the characteristic value by multiple fisrt feature parameters of the account to be identified
The the first machine sort model for training, obtains the first risk assessment value；The fisrt feature parameter is reflection
The parameter of the corresponding non-accumulative behavioural characteristic of the account；
Second processing module, for by the first risk assessment value, and at least the one of the account to be identified
The second machine sort model that the characteristic value input of second feature parameter is trained is planted, the second risk assessment is obtained
Value；The second feature parameter is the parameter of the corresponding accumulative behavioural characteristic of the reflection account；
Risk identification module, for based on the second risk assessment value for obtaining, the Business Processing is asked into
Row risk identification.
The embodiment of the present application is not the machine sort that all characteristic parameters of account are all directly inputted training
Model obtains risk assessment value, but first by the fisrt feature parameter of non-for various reflections accumulative behavioural characteristic
The first machine sort model that characteristic value input is trained, then the first wind that the first machine sort model is exported
Dangerous assessed value is input into the second machine sort model as a kind of characteristic value of new characteristic parameter, due to the first wind
Dangerous assessed value is the comprehensive assessment value obtained based on multiple fisrt feature parameters, such that it is able to weaken noise data
Impact, improve risk identification accuracy.
Description of the drawings
Fig. 1 is the Risk Identification Method flow chart that the embodiment of the present application one is provided；
Fig. 2 is the Risk Identification Method flow chart that the embodiment of the present application two is provided；
The risk identification apparatus structure schematic diagram that Fig. 3 is provided for the embodiment of the present application.
Specific embodiment
In the embodiment of the present application, server is receiving the business for account to be identified of user terminal transmission
After processing request, first by multiple fisrt feature parameters (corresponding non-accumulative behavior of reflection account of the account
The parameter of feature) characteristic value the first machine sort model for training of input, obtain the first risk assessment value；
Then, at least one second feature parameter (the reflection account by the first risk assessment value for obtaining, with the account
The parameter of the corresponding accumulative behavioural characteristic in family) characteristic value the second machine sort model for training of input, obtain
To the second risk assessment value；Based on the second risk assessment value for obtaining, the Business Processing of user terminal is asked
Carry out risk identification.
It can be seen that, the embodiment of the present application is not the machine that all characteristic parameters of account are all directly inputted training
Disaggregated model obtains risk assessment value, but trains the characteristic value input of various fisrt feature parameters first
The first machine sort model, then using the first machine sort model export the first risk assessment value as one kind
The characteristic value of new characteristic parameter is input into the second machine sort model, as the first risk assessment value is based on many
The comprehensive assessment value that fisrt feature parameter is obtained is planted, such that it is able to weaken the impact of noise data, risk is improved
The accuracy of identification.
The embodiment of the present application is described in further detail with reference to Figure of description.
Embodiment one
As shown in figure 1, the Risk Identification Method flow chart provided for the embodiment of the present application one, including following step
Suddenly：
S101：The Business Processing for account to be identified that server receive user terminal sends is asked.
Here Business Processing request can be information acquisition request, transaction request etc..Here account can be with
Refer to account (the member account such as registered) of the user terminal in server registration, it is also possible to refer to service
Device carries out the holding account (such as bank's card number) required for Business Processing for user terminal.
S102：First that the characteristic value input of multiple fisrt feature parameters of the account to be identified is trained
Machine sort model, obtains the first risk assessment value；The fisrt feature parameter is corresponding for the reflection account
Non- accumulative behavioural characteristic parameter.
Here, the first machine sort model can be specifically Logic Regression Models, it is of course also possible to according to instruction
Practice result need select other machine sort models.Here fisrt feature parameter is corresponding for reflection account
The parameter of non-accumulative behavioural characteristic, can include that conflict class parameter and initial parameter, conflict class parameter are referred to instead
The parameter with the presence or absence of conflict between different information is reflected, whether such as hair fastener state is conflicted with Billing Address state, is used
Whether the IP address of family terminal is conflicted with Billing Address state, characteristic parameter of the initial parameter for account itself,
Such as credit card issuer, Card Type (such as platinum card, gold card, generic card) etc..Characteristic value is to believe characteristic parameter
Breath carries out the value after processing that quantizes, and span can be [0,1].
Corresponding with the use of above-mentioned first machine sort model, when the first machine sort model is trained, need
(sample account here is the account of known Account Type to obtain each sample account in multiple sample accounts
Family) multiple fisrt feature parameters characteristic value, in conjunction with the Account Type of each sample account, train
One machine sort model.Such as, if a sample account is secured account, can be corresponding by secured account
Characteristic value (such as 0) as the first machine sort model target output value, by the every of the sample account
The characteristic value of individual fisrt feature parameter carries out model training as the input value of the first machine sort model.In detail
See the description of embodiment two.
S103：By the first risk assessment value, and at least one second feature ginseng of the account to be identified
The second machine sort model that several characteristic value inputs is trained, obtains the second risk assessment value；Described second
Characteristic parameter is the parameter of the corresponding accumulative behavioural characteristic of the reflection account.
In the step, the first risk assessment value that S102 is obtained is used as a kind of accumulative behavioural characteristic of reflection
The characteristic value of parameter, at least one corresponding with account itself is (generally multiple) to reflect accumulative behavioural characteristic
Parameter (change in such as one month and initiate Business Processing and ask the number of times of used IP address, default
Initiate number of times of Business Processing request etc. in time span) characteristic value together, be input into the second machine sort mould
Type, obtains the second risk assessment value.Second machine sort model can be specifically decision-tree model, gloomy at random
Woods model or other models.
Corresponding with the use of above-mentioned second machine sort model, when the second machine sort model is trained, need
Determine the corresponding first risk assessment value of each sample account in multiple sample accounts, and combine each sample
The Account Type of account, trains the second machine sort model.The description of detailed in Example two.
S104：Based on the second risk assessment value for obtaining, risk identification is carried out to Business Processing request.
In being embodied as, the second risk assessment value is bigger, can represent this Business Processing of user terminal
Ask the possibility for risk request bigger, a risk threshold value can be set, if the second risk assessment value is little
In the risk threshold value, be then that the account to be identified executes Business Processing, if the second risk assessment value more than or
The risk threshold value is equal to, then refusal executes Business Processing for the account to be identified.
Embodiment two
As shown in Fig. 2 the Risk Identification Method flow chart provided for the embodiment of the present application two, including following step
Suddenly：
S201：Server obtains the characteristic parameter information of the sample account of multiple known Account Types；Wherein,
The characteristic parameter information of each sample account includes the characteristic value of multiple fisrt feature parameters and at least one the
The characteristic value of two characteristic parameters；The Account Type includes secured account and adventure account, fisrt feature parameter
For reflecting the parameter of the corresponding non-accumulative behavioural characteristic of account, second feature parameter is corresponding tired for reflection account
The parameter of meter behavioural characteristic.
In specific implementation process, server can pass through complaint suggestion, service observation of user etc. and confirm to ask
The Account Type of the account of Business Processing was sought, credit rating is good, do not received related complaint suggestion
Account is classified as secured account, and credit rating is very poor, received the related account for complaining suggestion and is classified as adventure account.
For every kind of Account Type, the characteristic parameter information of multiple sample accounts can be obtained respectively, can wherein be wrapped
Include such as credit card issuer, Card Type, hair fastener state and Billing Address state the fisrt feature parameter such as whether to conflict, also include
As changed in one month in the number of times for initiating the used IP address of Business Processing request, predetermined time period
Initiate the second feature parameters such as the number of times of Business Processing request.
S202：Feature based on multiple fisrt feature parameters of each sample account in the plurality of sample account
Value, and the Account Type of each sample account, train the first machine sort model.
In specific implementation process, the first machine sort model can be trained, make the first machine sort model exist
Using the characteristic value of the fisrt feature parameter of multiple settings of each sample account as in the case of input value, defeated
Go out value and mate the risk assessment value corresponding to the Account Type of each sample account as far as possible.
Specifically, for each the sample account in the plurality of sample account, with the multiple of the sample account
Input value of the characteristic value of fisrt feature parameter as the first machine sort model, with the sample account
The corresponding risk assessment value of Account Type is trained as the target output value of the first machine sort model
The first machine sort model.
In specific implementation process, the first machine sort model can adopt Logic Regression Models, such as：
Wherein, S represents the output valve of Logic Regression Models, θiFor logistic regression coefficient, (as S202 needs
The coefficient of training), fiIt is the characteristic value under i-th kind of fisrt feature parameter, f0=1, n join for fisrt feature
Several kind numbers.
In being embodied as, in order to further reduce the complexity of Logic Regression Models, it is to avoid over-fitting occur,
Regular terms λ Ω (θ can be addedi) (λ is penalty coefficient), the regular terms is the monotonic increase of model complexity
Function, model are more complicated, and regular terms is bigger.Regular terms can take different forms, can such as make even
Side loses, and is exactly the L2 norms of parameter, it is also possible to take L1 norms.
It should be noted that using corresponding for the Account Type of sample account risk assessment value as the first machine point
The target output value of class model carries out model training, after the first machine sort model is trained, by each sample
Obtained by the first machine sort model that the characteristic value input of the various fisrt feature parameters of this account is trained
The first risk assessment value be not necessarily equal to carry out sample account during model training Account Type corresponding
Risk assessment value (i.e. target output value).
S203：Described that the input of the characteristic value of the various fisrt feature parameters of each sample account is trained
One machine sort model, obtains the first risk assessment value；It is based in the plurality of sample account, each sample
Account distinguishes characteristic value, the first risk assessment value and the account of corresponding various second feature parameters
Type, trains the second machine sort model.
In specific implementation process, the second machine sort model can be trained, make the second machine sort model exist
With the first risk assessment value of each sample account, and the characteristic value of the second feature parameter of multiple settings is made
In the case of for input value, the risk that output valve is tried one's best corresponding to the Account Type for mating each sample account is commented
Valuation.
Specifically, for each the sample account in the plurality of sample account, with the various of the sample account
The input of the characteristic value of second feature parameter and the first risk assessment value as the second machine sort model
Value, with the corresponding risk assessment value of the Account Type of the sample account, (account of such as security type is corresponding
Risk assessment value is 0, and the corresponding risk assessment value of the account of risk classifications is 1) as second machine
The target output value of disaggregated model, trains the second machine sort model.
Here, the second machine sort model can be the model pattern such as decision-tree model, Random Forest model.
S204：Server is asked in the Business Processing for account to be identified for receiving user terminal transmission
Afterwards, the first machine point for the characteristic value input of multiple fisrt feature parameters of the account to be identified being trained
Class model, obtains the first risk assessment value.
Here, multiple fisrt feature parameters of account to be identified, be and sample during the first machine sort model of training
Multiple fisrt feature parameter identical characteristic parameters of this account.Here the first risk assessment value is equivalent to right
The characteristic value of various fisrt feature parameters carries out the value after comprehensive refinement.
S205：By the first risk assessment value, and at least one second feature ginseng of the account to be identified
The second machine sort model that several characteristic value inputs is trained, obtains the second risk assessment value.
Here, at least one second feature parameter of account to be identified, be and the second machine sort model of training
When sample account at least one second feature parameter identical characteristic parameter.
S206：Based on the second risk assessment value for obtaining, risk identification is carried out to Business Processing request.
In being embodied as, if the second risk assessment value can respond user less than the risk threshold value for setting
In Business Processing request, be that account to be identified executes Business Processing, if the second risk assessment value more than or wait
In the risk threshold value for setting, then can refuse to execute Business Processing for the account to be identified.
Based on same inventive concept, additionally provide in the embodiment of the present application a kind of corresponding with Risk Identification Method
Risk identification device, due to principle and the embodiment of the present application Risk Identification Method phase of the device solve problem
Seemingly, therefore the enforcement of the device may refer to the enforcement of method, repeats part and repeats no more.
As shown in figure 3, the risk identification apparatus structure schematic diagram provided for the embodiment of the present application, including：
Receiver module 31, asks for the Business Processing for account to be identified that receive user terminal sends；
First processing module 32, for will be defeated for the characteristic value of multiple fisrt feature parameters of the account to be identified
Enter the first machine sort model for training, obtain the first risk assessment value；The fisrt feature parameter is anti-
Reflect the parameter of the corresponding non-accumulative behavioural characteristic of the account；
Second processing module 33, for by the first risk assessment value, and the account to be identified is at least
The second machine sort model that a kind of characteristic value input of second feature parameter is trained, obtains the second risk and comments
Valuation；The second feature parameter is the parameter of the corresponding accumulative behavioural characteristic of the reflection account；
Risk identification module 34, for based on the second risk assessment value for obtaining, asking to the Business Processing
Carry out risk identification.
Alternatively, described device also includes：
Model training module 35, believes for obtaining the characteristic parameter of the sample account of multiple known Account Types
Breath；Wherein, the characteristic parameter information of each sample account includes the characteristic value of multiple fisrt feature parameters；Institute
Stating Account Type includes secured account and adventure account；Based on each sample account in the plurality of sample account
Multiple fisrt feature parameters characteristic value, and the Account Type of each sample account trains described
One machine sort model.
Alternatively, the model training module 35 specifically for：
For each the sample account in the plurality of sample account, with multiple fisrt feature of the sample account
Input value of the characteristic value of parameter as the first machine sort model, with the Account Type of the sample account
Corresponding risk assessment value trains described first as the target output value of the first machine sort model
Machine sort model.
Alternatively, the characteristic parameter information of each sample account also includes the spy of at least one second feature parameter
Value indicative；The model training module 35 is additionally operable to：
First machine that the characteristic value input of the various fisrt feature parameters of each sample account is trained
Device disaggregated model, obtains the first risk assessment value；It is based in the plurality of sample account, each sample account
The characteristic value of the corresponding various second feature parameters of difference, the first risk assessment value and Account Type,
Train the second machine sort model.
Alternatively, the model training module 35 specifically for：
For each the sample account in the plurality of sample account, with the various second feature of the sample account
The input value of the characteristic value of parameter and the first risk assessment value as the second machine sort model, with the sample
The corresponding risk assessment value of the Account Type of this account is exported as the target of the second machine sort model
Value, trains the second machine sort model.
Those skilled in the art are it should be appreciated that embodiments herein can be provided as method, system or meter
Calculation machine program product.Therefore, the application can adopt complete hardware embodiment, complete software embodiment or knot
The form of the embodiment in terms of conjunction software and hardware.And, the application can be adopted and wherein be wrapped one or more
Computer-usable storage medium containing computer usable program code (including but not limited to magnetic disc store,
CD-ROM, optical memory etc.) the upper computer program that implements form.
The application is produced with reference to the method according to the embodiment of the present application, device (system) and computer program
The flow chart and/or block diagram of product is describing.It should be understood that can by computer program instructions flowchart and
/ or block diagram in each flow process and/or square frame and flow chart and/or the flow process in block diagram and/
Or the combination of square frame.These computer program instructions can be provided to all-purpose computer, special-purpose computer, embedded
The processor of formula processor or other programmable data processing devices is producing a machine so that by calculating
The instruction of the computing device of machine or other programmable data processing devices is produced for realization in flow chart one
The device of the function of specifying in individual flow process or one square frame of multiple flow processs and/or block diagram or multiple square frames.
These computer program instructions may be alternatively stored in and computer or the process of other programmable datas can be guided to set
In the standby computer-readable memory for working in a specific way so that be stored in the computer-readable memory
Instruction produce and include the manufacture of command device, command device realization is in one flow process or multiple of flow chart
The function of specifying in one square frame of flow process and/or block diagram or multiple square frames.
These computer program instructions can be also loaded in computer or other programmable data processing devices, made
Obtaining, series of operation steps is executed on computer or other programmable devices to produce computer implemented place
Reason, the instruction so as to execute on computer or other programmable devices are provided for realizing in flow chart one
The step of function of specifying in flow process or one square frame of multiple flow processs and/or block diagram or multiple square frames.
Although having been described for the preferred embodiment of the application, those skilled in the art once know base
This creative concept, then can make other change and modification to these embodiments.So, appended right will
Ask and be intended to be construed to include preferred embodiment and fall into the had altered of the application scope and change.
Obviously, those skilled in the art can carry out various changes and modification without deviating from this Shen to the application
Spirit and scope please.So, if the application these modification and modification belong to the application claim and
Within the scope of its equivalent technologies, then the application is also intended to comprising these changes and modification.
Claims (10)
1. a kind of Risk Identification Method, it is characterised in that the method includes：
The Business Processing for account to be identified that server receive user terminal sends is asked；
The first machine that the characteristic value input of multiple fisrt feature parameters of the account to be identified is trained
Disaggregated model, obtains the first risk assessment value；The fisrt feature parameter is corresponding non-for the reflection account
The parameter of accumulative behavioural characteristic；
By the first risk assessment value, and the spy of at least one second feature parameter of the account to be identified
The second machine sort model that value indicative input is trained, obtains the second risk assessment value；The second feature ginseng
Parameter of the number for the corresponding accumulative behavioural characteristic of the reflection account；
Based on the second risk assessment value for obtaining, risk identification is carried out to Business Processing request.
2. the method for claim 1, it is characterised in that train described according to following steps
One machine sort model：
Obtain the characteristic parameter information of the sample account of multiple known Account Types；Wherein, each sample account
Characteristic parameter information include the characteristic value of multiple fisrt feature parameters；The Account Type includes secured account
And adventure account；
Based on the characteristic value of multiple fisrt feature parameters of each sample account in the plurality of sample account, with
And the Account Type of each sample account, train the first machine sort model.
3. method as claimed in claim 2, it is characterised in that based in the plurality of sample account per
The characteristic value of multiple fisrt feature parameters of individual sample account, and the Account Type of each sample account, instruction
The first machine sort model is practised, including：
For each the sample account in the plurality of sample account, with multiple fisrt feature of the sample account
Input value of the characteristic value of parameter as the first machine sort model, with the Account Type of the sample account
Corresponding risk assessment value trains described first as the target output value of the first machine sort model
Machine sort model.
4. method as claimed in claim 2 or claim 3, it is characterised in that the feature ginseng of each sample account
Number information also includes the characteristic value of at least one second feature parameter；Described second is trained according to following steps
Machine sort model：
First machine that the characteristic value input of the various fisrt feature parameters of each sample account is trained
Device disaggregated model, obtains the first risk assessment value；
Based in the plurality of sample account, each sample account distinguishes corresponding various second feature parameters
Characteristic value, the first risk assessment value and Account Type, train the second machine sort model.
5. method as claimed in claim 4, it is characterised in that in based on the plurality of sample account,
Each sample account distinguish the characteristic value of corresponding various second feature parameters, the first risk assessment value,
And Account Type, the second machine sort model is trained, including：
For each the sample account in the plurality of sample account, with the various second feature of the sample account
The input value of the characteristic value of parameter and the first risk assessment value as the second machine sort model, with the sample
The corresponding risk assessment value of the Account Type of this account is exported as the target of the second machine sort model
Value, trains the second machine sort model.
6. a kind of risk identification device, it is characterised in that the device includes：
Receiver module, asks for the Business Processing for account to be identified that receive user terminal sends；
First processing module, is input into for the characteristic value by multiple fisrt feature parameters of the account to be identified
The the first machine sort model for training, obtains the first risk assessment value；The fisrt feature parameter is reflection
The parameter of the corresponding non-accumulative behavioural characteristic of the account；
Second processing module, for by the first risk assessment value, and at least the one of the account to be identified
The second machine sort model that the characteristic value input of second feature parameter is trained is planted, the second risk assessment is obtained
Value；The second feature parameter is the parameter of the corresponding accumulative behavioural characteristic of the reflection account；
Risk identification module, for based on the second risk assessment value for obtaining, the Business Processing is asked into
Row risk identification.
7. device as claimed in claim 6, it is characterised in that described device also includes：
Model training module, for obtaining the characteristic parameter information of the sample account of multiple known Account Types；
Wherein, the characteristic parameter information of each sample account includes the characteristic value of multiple fisrt feature parameters；The account
Family type includes secured account and adventure account；Based in the plurality of sample account each sample account many
The characteristic value of fisrt feature parameter, and the Account Type of each sample account is planted, first machine is trained
Device disaggregated model.
8. device as claimed in claim 7, it is characterised in that the model training module specifically for：
For each the sample account in the plurality of sample account, with multiple fisrt feature of the sample account
Input value of the characteristic value of parameter as the first machine sort model, with the Account Type of the sample account
Corresponding risk assessment value trains described first as the target output value of the first machine sort model
Machine sort model.
9. device as claimed in claim 7 or 8, it is characterised in that the feature ginseng of each sample account
Number information also includes the characteristic value of at least one second feature parameter；The model training module is additionally operable to：
First machine that the characteristic value input of the various fisrt feature parameters of each sample account is trained
Device disaggregated model, obtains the first risk assessment value；It is based in the plurality of sample account, each sample account
The characteristic value of the corresponding various second feature parameters of difference, the first risk assessment value and Account Type,
Train the second machine sort model.
10. device as claimed in claim 9, it is characterised in that the model training module specifically for：
For each the sample account in the plurality of sample account, with the various second feature of the sample account
The input value of the characteristic value of parameter and the first risk assessment value as the second machine sort model, with the sample
The corresponding risk assessment value of the Account Type of this account is exported as the target of the second machine sort model
Value, trains the second machine sort model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510560860.5A CN106503562A (en) | 2015-09-06 | 2015-09-06 | A kind of Risk Identification Method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510560860.5A CN106503562A (en) | 2015-09-06 | 2015-09-06 | A kind of Risk Identification Method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106503562A true CN106503562A (en) | 2017-03-15 |
Family
ID=58286292
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510560860.5A Pending CN106503562A (en) | 2015-09-06 | 2015-09-06 | A kind of Risk Identification Method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106503562A (en) |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107391569A (en) * | 2017-06-16 | 2017-11-24 | 阿里巴巴集团控股有限公司 | Identification, model training, Risk Identification Method, device and the equipment of data type |
| CN107423883A (en) * | 2017-06-15 | 2017-12-01 | 阿里巴巴集团控股有限公司 | Risk Identification Method and device, the electronic equipment of pending business |
| CN107609972A (en) * | 2017-07-31 | 2018-01-19 | 阿里巴巴集团控股有限公司 | A kind of risk business recognition method and device |
| WO2018184520A1 (en) * | 2017-04-07 | 2018-10-11 | 平安科技（深圳）有限公司 | Service data processing method, device, server and storage medium |
| CN108876076A (en) * | 2017-05-09 | 2018-11-23 | 中国移动通信集团广东有限公司 | The personal credit methods of marking and device of data based on instruction |
| CN109510800A (en) * | 2017-09-14 | 2019-03-22 | 北京金山云网络技术有限公司 | A kind of network request processing method, device, electronic equipment and storage medium |
| CN109544166A (en) * | 2018-11-05 | 2019-03-29 | 阿里巴巴集团控股有限公司 | A kind of Risk Identification Method and device |
| CN109657696A (en) * | 2018-11-05 | 2019-04-19 | 阿里巴巴集团控股有限公司 | Multitask supervised learning model training, prediction technique and device |
| CN109714301A (en) * | 2017-10-25 | 2019-05-03 | 北京京东尚科信息技术有限公司 | Register Risk Identification Method, device, electronic equipment and storage medium |
| CN109767337A (en) * | 2017-11-09 | 2019-05-17 | 腾讯科技（深圳）有限公司 | The recognition methods of adverse selection user, device and computer equipment in insurance |
| CN109766496A (en) * | 2018-12-28 | 2019-05-17 | 北京奇安信科技有限公司 | A kind of content risks recognition methods, system, equipment and medium |
| CN109919783A (en) * | 2019-01-31 | 2019-06-21 | 德联易控科技(北京)有限公司 | Risk Identification Method, device, equipment and the storage medium of vehicle insurance Claims Resolution case |
| CN110245959A (en) * | 2019-04-17 | 2019-09-17 | 阿里巴巴集团控股有限公司 | The treating method and apparatus of specific aim request |
| WO2019174393A1 (en) * | 2018-03-14 | 2019-09-19 | 阿里巴巴集团控股有限公司 | Graph structure model training and junk account identification |
| TWI679604B (en) * | 2017-05-26 | 2019-12-11 | 香港商阿里巴巴集團服務有限公司 | Method and device for determining user risk level, computer equipment |
| CN110580170A (en) * | 2019-09-12 | 2019-12-17 | 中国工商银行股份有限公司 | software performance risk identification method and device |
| CN111125695A (en) * | 2019-12-26 | 2020-05-08 | 武汉极意网络科技有限公司 | Account risk assessment method, device, equipment and storage medium |
| CN111260368A (en) * | 2020-01-08 | 2020-06-09 | 支付宝(杭州)信息技术有限公司 | Account transaction risk judgment method and device and electronic equipment |
| CN111275507A (en) * | 2018-12-04 | 2020-06-12 | 北京嘀嘀无限科技发展有限公司 | Order abnormity identification and order risk management and control method and system |
| CN111275470A (en) * | 2018-12-04 | 2020-06-12 | 北京嘀嘀无限科技发展有限公司 | Service initiation probability prediction method and training method and device of model thereof |
| US10783457B2 (en) | 2017-05-26 | 2020-09-22 | Alibaba Group Holding Limited | Method for determining risk preference of user, information recommendation method, and apparatus |
| CN112801800A (en) * | 2021-04-14 | 2021-05-14 | 深圳格隆汇信息科技有限公司 | Behavior fund analysis system, behavior fund analysis method, computer equipment and storage medium |
| CN112905987A (en) * | 2019-11-19 | 2021-06-04 | 北京达佳互联信息技术有限公司 | Account identification method, account identification device, server and storage medium |
| CN114157490A (en) * | 2021-12-03 | 2022-03-08 | 武汉极意网络科技有限公司 | User request event analysis method based on clustering algorithm |
-
2015
- 2015-09-06 CN CN201510560860.5A patent/CN106503562A/en active Pending
Cited By (40)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018184520A1 (en) * | 2017-04-07 | 2018-10-11 | 平安科技（深圳）有限公司 | Service data processing method, device, server and storage medium |
| CN108876076A (en) * | 2017-05-09 | 2018-11-23 | 中国移动通信集团广东有限公司 | The personal credit methods of marking and device of data based on instruction |
| US10783457B2 (en) | 2017-05-26 | 2020-09-22 | Alibaba Group Holding Limited | Method for determining risk preference of user, information recommendation method, and apparatus |
| TWI679604B (en) * | 2017-05-26 | 2019-12-11 | 香港商阿里巴巴集團服務有限公司 | Method and device for determining user risk level, computer equipment |
| US11494774B2 (en) | 2017-06-15 | 2022-11-08 | Advanced New Technologies Co., Ltd. | Method, apparatus and electronic device for identifying risks pertaining to transactions to be processed |
| CN107423883A (en) * | 2017-06-15 | 2017-12-01 | 阿里巴巴集团控股有限公司 | Risk Identification Method and device, the electronic equipment of pending business |
| CN107423883B (en) * | 2017-06-15 | 2020-04-07 | 创新先进技术有限公司 | Risk identification method and device for to-be-processed service and electronic equipment |
| US11367075B2 (en) | 2017-06-15 | 2022-06-21 | Advanced New Technologies Co., Ltd. | Method, apparatus and electronic device for identifying risks pertaining to transactions to be processed |
| US11113394B2 (en) | 2017-06-16 | 2021-09-07 | Advanced New Technologies Co., Ltd. | Data type recognition, model training and risk recognition methods, apparatuses and devices |
| CN107391569B (en) * | 2017-06-16 | 2020-09-15 | 阿里巴巴集团控股有限公司 | Data type identification, model training and risk identification method, device and equipment |
| CN107391569A (en) * | 2017-06-16 | 2017-11-24 | 阿里巴巴集团控股有限公司 | Identification, model training, Risk Identification Method, device and the equipment of data type |
| US11100220B2 (en) | 2017-06-16 | 2021-08-24 | Advanced New Technologies Co., Ltd. | Data type recognition, model training and risk recognition methods, apparatuses and devices |
| CN107609972A (en) * | 2017-07-31 | 2018-01-19 | 阿里巴巴集团控股有限公司 | A kind of risk business recognition method and device |
| CN109510800B (en) * | 2017-09-14 | 2020-11-27 | 北京金山云网络技术有限公司 | Network request processing method and device, electronic equipment and storage medium |
| CN109510800A (en) * | 2017-09-14 | 2019-03-22 | 北京金山云网络技术有限公司 | A kind of network request processing method, device, electronic equipment and storage medium |
| CN109714301A (en) * | 2017-10-25 | 2019-05-03 | 北京京东尚科信息技术有限公司 | Register Risk Identification Method, device, electronic equipment and storage medium |
| CN109714301B (en) * | 2017-10-25 | 2021-11-30 | 北京京东尚科信息技术有限公司 | Registration risk identification method and device, electronic equipment and storage medium |
| CN109767337A (en) * | 2017-11-09 | 2019-05-17 | 腾讯科技（深圳）有限公司 | The recognition methods of adverse selection user, device and computer equipment in insurance |
| US10917425B2 (en) | 2018-03-14 | 2021-02-09 | Advanced New Technologies Co., Ltd. | Graph structure model training and junk account identification |
| WO2019174393A1 (en) * | 2018-03-14 | 2019-09-19 | 阿里巴巴集团控股有限公司 | Graph structure model training and junk account identification |
| CN109544166A (en) * | 2018-11-05 | 2019-03-29 | 阿里巴巴集团控股有限公司 | A kind of Risk Identification Method and device |
| CN109657696A (en) * | 2018-11-05 | 2019-04-19 | 阿里巴巴集团控股有限公司 | Multitask supervised learning model training, prediction technique and device |
| CN109544166B (en) * | 2018-11-05 | 2023-05-30 | 创新先进技术有限公司 | Risk identification method and risk identification device |
| CN109657696B (en) * | 2018-11-05 | 2023-06-30 | 创新先进技术有限公司 | Multi-task supervised learning model training and predicting method and device |
| CN111275470A (en) * | 2018-12-04 | 2020-06-12 | 北京嘀嘀无限科技发展有限公司 | Service initiation probability prediction method and training method and device of model thereof |
| CN111275470B (en) * | 2018-12-04 | 2023-12-01 | 北京嘀嘀无限科技发展有限公司 | Service initiation probability prediction method and training method and device of model thereof |
| CN111275507A (en) * | 2018-12-04 | 2020-06-12 | 北京嘀嘀无限科技发展有限公司 | Order abnormity identification and order risk management and control method and system |
| CN109766496B (en) * | 2018-12-28 | 2021-02-09 | 奇安信科技集团股份有限公司 | Content risk identification method, system, device and medium |
| CN109766496A (en) * | 2018-12-28 | 2019-05-17 | 北京奇安信科技有限公司 | A kind of content risks recognition methods, system, equipment and medium |
| CN109919783A (en) * | 2019-01-31 | 2019-06-21 | 德联易控科技(北京)有限公司 | Risk Identification Method, device, equipment and the storage medium of vehicle insurance Claims Resolution case |
| CN110245959A (en) * | 2019-04-17 | 2019-09-17 | 阿里巴巴集团控股有限公司 | The treating method and apparatus of specific aim request |
| CN110580170A (en) * | 2019-09-12 | 2019-12-17 | 中国工商银行股份有限公司 | software performance risk identification method and device |
| CN110580170B (en) * | 2019-09-12 | 2023-07-21 | 中国工商银行股份有限公司 | Method and device for identifying software performance risk |
| CN112905987A (en) * | 2019-11-19 | 2021-06-04 | 北京达佳互联信息技术有限公司 | Account identification method, account identification device, server and storage medium |
| CN112905987B (en) * | 2019-11-19 | 2024-02-27 | 北京达佳互联信息技术有限公司 | Account identification method, device, server and storage medium |
| CN111125695B (en) * | 2019-12-26 | 2022-04-05 | 武汉极意网络科技有限公司 | Account risk assessment method, device, equipment and storage medium |
| CN111125695A (en) * | 2019-12-26 | 2020-05-08 | 武汉极意网络科技有限公司 | Account risk assessment method, device, equipment and storage medium |
| CN111260368A (en) * | 2020-01-08 | 2020-06-09 | 支付宝(杭州)信息技术有限公司 | Account transaction risk judgment method and device and electronic equipment |
| CN112801800A (en) * | 2021-04-14 | 2021-05-14 | 深圳格隆汇信息科技有限公司 | Behavior fund analysis system, behavior fund analysis method, computer equipment and storage medium |
| CN114157490A (en) * | 2021-12-03 | 2022-03-08 | 武汉极意网络科技有限公司 | User request event analysis method based on clustering algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106503562A (en) | A kind of Risk Identification Method and device | |
| US10115108B1 (en) | Rendering transaction data to identify fraud detection rule strength | |
| US11900271B2 (en) | Self learning data loading optimization for a rule engine | |
| US11798017B2 (en) | Dynamic information probing for classifying an item | |
| CN107563757B (en) | Data risk identification method and device | |
| CN108665366A (en) | Determine method, terminal device and the computer readable storage medium of consumer's risk grade | |
| US20140258119A1 (en) | Fraud Decision Processing System and Method | |
| TWI752349B (en) | Risk identification method and device | |
| CN108765144B (en) | Electronic resource package processing method, terminal device and medium | |
| WO2020238235A1 (en) | Identity verification and checking method and apparatus | |
| CN108038687A (en) | Method of commerce, server and computer-readable recording medium based on speech recognition | |
| CN110349013A (en) | Risk control method and device | |
| KR20190130535A (en) | Pair Trading System based on the Stock bigdata and Controlling Method for the Same | |
| US20190362241A1 (en) | Systems and methods for configuring an online decision engine | |
| US11786825B2 (en) | Fraud detection system | |
| CN111160737A (en) | Adaptation method of resource allocation scheme and related equipment | |
| CN110162957B (en) | Authentication method and device for intelligent equipment, storage medium and electronic device | |
| CN110852809A (en) | Data processing method, device, equipment and medium | |
| CN107679862B (en) | Method and device for determining characteristic value of fraud transaction model | |
| CN112330373A (en) | User behavior analysis method and device and computer readable storage medium | |
| CN112330355A (en) | Consumption ticket transaction data processing method, device, equipment and storage medium | |
| CN111967948A (en) | Bank product recommendation method and device, server and storage medium | |
| CN114971638A (en) | Transaction authentication method and device based on risk identification | |
| US20220036219A1 (en) | Systems and methods for fraud detection using game theory | |
| CN110427971A (en) | Recognition methods, device, server and the storage medium of user and IP |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200922 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Innovative advanced technology Co.,Ltd. Address before: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant before: Advanced innovation technology Co.,Ltd. Effective date of registration: 20200922 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Advanced innovation technology Co.,Ltd. Address before: Cayman Islands Grand Cayman capital building, a four storey No. 847 mailbox Applicant before: Alibaba Group Holding Ltd. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170315 |