CN107240024A - The anti-fraud recognition methods of settlement of insurance claim and device - Google Patents
The anti-fraud recognition methods of settlement of insurance claim and device Download PDFInfo
- Publication number
- CN107240024A CN107240024A CN201710365366.2A CN201710365366A CN107240024A CN 107240024 A CN107240024 A CN 107240024A CN 201710365366 A CN201710365366 A CN 201710365366A CN 107240024 A CN107240024 A CN 107240024A
- Authority
- CN
- China
- Prior art keywords
- resolution
- fraud
- identification model
- risk
- case 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
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
- G06Q40/08—Insurance
-
- 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
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
- G06Q30/0185—Product, service or business identity fraud
Abstract
Anti- fraud recognition methods and device the present invention relates to a kind of settlement of insurance claim, this method include:Anti- fake system settle a claim after the Claims Resolution application for carrying Claims Resolution information is received, it is determined that the corresponding Claims Resolution case type of Claims Resolution application, obtains the anti-fraud identification model of the corresponding training in advance generation of Claims Resolution application;The risks and assumptions of the corresponding anti-fraud identification model of Claims Resolution application are obtained, the information of the risks and assumptions of the corresponding anti-fraud identification model of Claims Resolution application is obtained from predetermined financial system;The information of risks and assumptions is inputted into the anti-fraud identification model, the corresponding risk indicator value of output Claims Resolution application;The corresponding risk processing mode of Claims Resolution application is determined based on risk indicator value and default risk processing rule, and risk processing is carried out to Claims Resolution application based on risk processing mode.The present invention can rapidly carry out the processing of the Claims Resolution application of batch, improve the efficiency of anti-fraud identification, and can recognize implicit risk of fraud to the full extent.
Description
Technical field
The present invention relates to banking and insurance business technical field, more particularly to a kind of anti-fraud recognition methods of settlement of insurance claim and dress
Put.
Background technology
At present, banking and insurance business field (such as in life insurance field), some clients handle related insurance Claims Resolution business when,
It can be insured for the purpose of insurance fraud, or provide false Claims Resolution information (for example, fabricating insurance risk, exaggerating insurance risk
Order of severity etc.).In order to evade this kind of risk, banking and insurance business relevant enterprise would generally set special Claims Review post, adopt
Claims Resolution application of the mode manually audited to client is based on professional experiences and carries out anti-fraud identification, and anti-fraud recognition efficiency is low,
And because the professional standards of each auditor differ, it is thus possible to there is part case not enough to the identification dynamics for implying risk.
The content of the invention
Anti- fraud recognition methods and device it is an object of the invention to provide a kind of settlement of insurance claim, it is intended to known with anti-fraud
Other model carries out uniform auditing to Claims Resolution application, improves the efficiency of anti-fraud identification, strengthens the dynamics of implicit risk identification.
To achieve the above object, the present invention provides a kind of anti-fraud recognition methods of settlement of insurance claim, the settlement of insurance claim
Anti- fraud recognition methods includes:
S1, settles a claim anti-fake system after the Claims Resolution application for carrying Claims Resolution information is received, and determines the Claims Resolution application pair
The Claims Resolution case type answered, the Claims Resolution is obtained according to predetermined Claims Resolution case type and the incidence relation of anti-fraud identification model
Apply for the anti-fraud identification model of corresponding training in advance generation;
S2, obtains the Claims Resolution application corresponding based on predetermined anti-fraud identification model and the incidence relation of risks and assumptions
The risks and assumptions of anti-fraud identification model, obtain the corresponding anti-fraud identification mould of the Claims Resolution application from predetermined financial system
The information of the risks and assumptions of type;
S3, the Claims Resolution is applied for the information of the risks and assumptions of corresponding anti-fraud identification model is inputted to the anti-fraud and known
In other model, corresponding risk indicator value is applied for export the Claims Resolution by the anti-fraud identification model;
S4, is determined at the corresponding risk of the Claims Resolution application based on the risk indicator value and default risk processing rule
Reason mode, and risk processing is carried out to the Claims Resolution application based on the risk processing mode.
Preferably, the Claims Resolution case type includes unexpected die class Claims Resolution case, disease and die class Claims Resolution case, great
Disease class Claims Resolution case and unexpected medical class Claims Resolution case.
Preferably, the default risk processing rule is:
If the Claims Resolution applies for that corresponding risk indicator value is more than default risk indicator threshold value, and risk indicator value is big
The case of artificial anti-fraud identifying processing is applied in the Claims Resolution of default risk indicator threshold value;
If the Claims Resolution applies for that corresponding risk indicator value is less than or equal to default risk indicator threshold value and more than or equal to pre-
If secure threshold, then the Claims Resolution application to predetermined number is inspected by random samples in default sampling observation ratio, with the Claims Resolution Shen inspected by random samples
It please be used as the artificial anti-case for cheating identifying processing.
Preferably, the anti-fraud identification model includes before being Logic Regression Models, the step S1:
S01, according to predetermined Claims Resolution case type and the mapping relations of the full dose factor, determines the Claims Resolution case type of selection
The corresponding full dose factor as sample data set variable;
S02, is the sample data that each Claims Resolution case type prepares predetermined number, is that each sample data demarcates corresponding wind
Dangerous desired value, forms sample data set;
S03, extracts the letter of identified each full dose factor from the corresponding sample data of Claims Resolution case type of selection
Breath;
S04, is divided into the first ratio by the information of the full dose factor of the corresponding sample data of Claims Resolution case type of selection
The checking collection of training set, the second ratio;
S05, trains corresponding Logic Regression Models using the corresponding training set of the Claims Resolution case type of selection, draws the reason
Pay for the incidence relation of the corresponding Logic Regression Models of case type and risks and assumptions;
S06, verifies the hit rate of the Logic Regression Models using the corresponding checking collection of the Claims Resolution case type of selection, covers
Lid rate, accuracy rate, if the hit rate, coverage rate and accuracy rate respectively correspond to more than or equal to default hit rate, coverage rate and
Accuracy rate, then training terminates, otherwise, increases the quantity of the sample data of training set and is trained again.
Preferably, the step S04 further comprises:By the full dose of the corresponding sample data of Claims Resolution case type of selection
The information of the factor is divided into the training set of the first ratio, the checking collection and the test set of the 3rd ratio of the second ratio, the step S06
Also include afterwards:
The Logic Regression Models that training terminates are tested using the corresponding test set of Claims Resolution case type of selection, and
Generate test report.
To achieve the above object, the present invention also provides a kind of anti-fraud identifying device of settlement of insurance claim, the settlement of insurance claim
Anti- fraud identifying device include:
First determining module, for after the Claims Resolution application for carrying Claims Resolution information is received, determining the Claims Resolution application pair
The Claims Resolution case type answered, the Claims Resolution is obtained according to predetermined Claims Resolution case type and the incidence relation of anti-fraud identification model
Apply for the anti-fraud identification model of corresponding training in advance generation;
Acquisition module, the Claims Resolution is obtained for the incidence relation based on predetermined anti-fraud identification model and risks and assumptions
Apply for the risks and assumptions of corresponding anti-fraud identification model, the Claims Resolution application is obtained from predetermined financial system corresponding anti-
Cheat the information of the risks and assumptions of identification model;
Output module, for by it is described Claims Resolution apply it is corresponding it is anti-fraud identification model risks and assumptions information input to
In the anti-fraud identification model, corresponding risk indicator value is applied for export the Claims Resolution by the anti-fraud identification model;
Processing module, for determining the Claims Resolution application pair based on the risk indicator value and default risk processing rule
The risk processing mode answered, and risk processing is carried out to the Claims Resolution application based on the risk processing mode.
Preferably, the Claims Resolution case type includes unexpected die class Claims Resolution case, disease and die class Claims Resolution case, great
Disease class Claims Resolution case and unexpected medical class Claims Resolution case.
Preferably, the default risk processing rule is:
If the Claims Resolution applies for that corresponding risk indicator value is more than default risk indicator threshold value, and risk indicator value is big
The case of artificial anti-fraud identifying processing is applied in the Claims Resolution of default risk indicator threshold value;If the Claims Resolution application correspondence
Risk indicator value be less than or equal to default risk indicator threshold value and more than or equal to default secure threshold, then to predetermined number
Claims Resolution application is inspected by random samples in default sampling observation ratio, and the case of artificial anti-fraud identifying processing is applied for the Claims Resolution inspected by random samples
Part.
Preferably, the processing unit of the settlement of insurance claim also includes:
Second determining module, for according to predetermined Claims Resolution case type and the mapping relations of the full dose factor, it is determined that selection
The corresponding full dose factor of Claims Resolution case type as sample data set variable;
Demarcating module, the sample data for preparing predetermined number for each Claims Resolution case type, is each sample data mark
Fixed corresponding risk indicator value, forms sample data set;
Extraction module, for extracting identified each full dose from the corresponding sample data of Claims Resolution case type of selection
The information of the factor;
Division module, for the information of the full dose factor of the corresponding sample data of Claims Resolution case type of selection to be divided into
The checking collection of the training set of one ratio, the second ratio;
Training module, corresponding logistic regression mould is trained for the corresponding training set of Claims Resolution case type using selection
Type, draws the incidence relation of the corresponding Logic Regression Models of Claims Resolution case type and risks and assumptions;
Authentication module, for the Claims Resolution case type corresponding checking collection checking Logic Regression Models using selection
Hit rate, coverage rate, accuracy rate, if the hit rate, coverage rate and accuracy rate respectively correspond to more than or equal to default hit rate,
Coverage rate and accuracy rate, then training terminate, otherwise, increase the quantity of the sample data of training set and are trained again.
Preferably, the division module is further used for the full dose of the corresponding sample data of Claims Resolution case type of selection
The information of the factor is divided into the training set of the first ratio, the checking collection and the test set of the 3rd ratio of the second ratio;
The processing unit of the settlement of insurance claim also includes:Test module, for making to the Logic Regression Models that training terminates
Tested with the corresponding test set of Claims Resolution case type of selection, and generate test report.
The beneficial effects of the invention are as follows:Because anti-fraud identification model is is trained generation using big data in advance,
Therefore, it is possible to recognize implicit risk of fraud to the full extent, strengthen the dynamics of implicit risk identification;In addition, using anti-fraud
Identification model obtains risk indicator value, to handle Claims Resolution application, can rapidly carry out the place of the Claims Resolution application of batch
Reason, improves the efficiency of anti-fraud identification.
Brief description of the drawings
Fig. 1 is the schematic flow sheet of the anti-fraud recognition methods first embodiment of settlement of insurance claim of the present invention;
Fig. 2 is the schematic flow sheet of the anti-fraud recognition methods second embodiment of settlement of insurance claim of the present invention;
Fig. 3 is the structural representation of the anti-fraud identifying device first embodiment of settlement of insurance claim of the present invention;
Fig. 4 is the structural representation of the anti-fraud identifying device second embodiment of settlement of insurance claim of the present invention.
Embodiment
The principle and feature of the present invention are described below in conjunction with accompanying drawing, the given examples are served only to explain the present invention, and
It is non-to be used to limit the scope of the present invention.
As shown in figure 1, schematic flow sheets of the Fig. 1 for the anti-embodiment of fraud recognition methods one of settlement of insurance claim of the present invention, should
The anti-fraud recognition methods of settlement of insurance claim comprises the following steps:
Step S1, settles a claim anti-fake system after the Claims Resolution application for carrying Claims Resolution information is received, determines the Claims Resolution Shen
Please corresponding Claims Resolution case type, according to being obtained the incidence relation of predetermined Claims Resolution case type and anti-fraud identification model
The anti-fraud identification model of the corresponding training in advance generation of Claims Resolution application;
In the present embodiment, the anti-fake system of settling a claim is installed in the anti-fraud identifying device of settlement of insurance claim, and client can be
Operated in the anti-fraud identifying device of settlement of insurance claim, log in the anti-fake system of Claims Resolution, to initiate Claims Resolution application, insurance reason
The anti-fraud identifying device paid for can be any suitable equipment such as mobile phone or computer.
Anti- fake system settle a claim after the Claims Resolution application of client is received, according to the Claims Resolution information carried in Claims Resolution application
(such as case type information of settling a claim, the customer information of throwing warrantee) determines the corresponding Claims Resolution case type of the Claims Resolution application,
Claims Resolution case type includes but is not limited to unexpected die class Claims Resolution case, disease and die class Claims Resolution case, major disease class reason
Pay for case and unexpected medical class Claims Resolution case.
In the present embodiment, training in advance generates multiple anti-fraud identification models and preserved in anti-fake system of settling a claim, counter to take advantage of
Cheating identification model is, for example,:The first anti-fraud identification model corresponding with unexpected class Claims Resolution case of dieing, class reason of being die with disease
Pay for the corresponding second anti-fraud identification model of case, identification model is instead cheated with major disease class Claims Resolution case the corresponding 3rd with
And instead cheat identification model with unexpected medical class Claims Resolution case the corresponding 4th.Each Claims Resolution case type association correspondence one is counter to take advantage of
Preserved in swindleness identification model, anti-fake system of settling a claim and pay for case type pass corresponding with the association one by one of anti-fraud identification model
System, it is determined that after the corresponding Claims Resolution case type of the Claims Resolution application, the corresponding training in advance generation of Claims Resolution application can be obtained
Anti- fraud identification model.
Preferably, with reference to Fig. 2, the training process of anti-fraud identification model includes:
S01, according to predetermined Claims Resolution case type and the mapping relations of the full dose factor, determines the Claims Resolution case type of selection
The corresponding full dose factor as sample data set variable;The full dose factor is the reason with corresponding Claims Resolution application prepared for modeling
The related all data factors of case type are paid for, risks and assumptions are the subset of the full dose factor.
S02, is the sample data (such as 100,000 parts sample datas) that each Claims Resolution case type prepares predetermined number, is each
Sample data demarcates corresponding risk indicator value, forms sample data set.
S03, extracts the letter of identified each full dose factor from the corresponding sample data of Claims Resolution case type of selection
Breath;The customer information that throwing insurant is for example extracted from the corresponding sample data of Claims Resolution case type of selection (is for example wrapped
Include name, age, annual income information etc.), disease and treatment (for example suffer from which kind of disease, the duration treated and control
Expense for the treatment of etc.), be in danger duration (be, for example, 1 month), insurer's loan profile of Claims Resolution case (for example include without loan, borrow
Money date and the amount of money etc.) and overdue information (for example, provide a loan and do not refund overdue 5 days) etc..
S04, is divided into the first ratio by the information of the full dose factor of the corresponding sample data of Claims Resolution case type of selection
The checking collection of training set, the second ratio, such as the corresponding sample data of each Claims Resolution case type, with the sample of 50% ratio
Notebook data collects as training set using the sample data of 25% ratio as checking.
S05, trains corresponding Logic Regression Models using the corresponding training set of the Claims Resolution case type of selection, draws the reason
Pay for the incidence relation of the corresponding Logic Regression Models of case type and risks and assumptions;Wherein, when training first, logistic regression mould
The parameter of type is the parameter of acquiescence, and with the progress of training, parameter is constantly adjusted;Risks and assumptions are corresponding logistic regression
The full dose factor of the specific part of model selection;In addition, in the training process, different full dose factor pair risk indicator values have not
With the influence of degree, can using some larger full dose factors of influence degree as risks and assumptions, then by these risks and assumptions with
Logic Regression Models are associated.
S06, verifies the hit rate of the Logic Regression Models using the corresponding checking collection of the Claims Resolution case type of selection, covers
Lid rate, accuracy rate, if the hit rate, coverage rate and accuracy rate respectively correspond to more than or equal to default hit rate, coverage rate and
Accuracy rate, default hit rate, coverage rate and accuracy rate are for example respectively 0.98,0.99 and 0.98, then training terminates;If tested
Demonstrate,prove covering of the hit rate less than default hit rate, or the gained of verifying logic regression model of the gained of Logic Regression Models
Rate is less than default coverage rate, or the accuracy rate of the gained of verifying logic regression model is less than default accuracy rate, then needs
Increase the quantity of the sample data of training set, and be trained again, until after training terminates, verified obtained by hit rate,
Coverage rate and accuracy rate are corresponded to respectively is more than or equal to default hit rate, coverage rate and accuracy rate, is obtained after being terminated with training
Logic Regression Models are used as anti-fraud identification model to be used.
In addition, anti-fraud identification model is preferably that (Logic Regression Models are in identification Claims Resolution application for Logic Regression Models
Stronger, application scenarios are wider in terms of the interpretation of risk of fraud), certainly anti-fraud identification model can also be other moulds
Type, such as neural network model, decision-tree model, excessive restriction is not done herein.
Preferably, above-mentioned steps S04 further comprises:By the full dose of the corresponding sample data of Claims Resolution case type of selection
The information of the factor is divided into the training set of the first ratio, the checking collection and the test set of the 3rd ratio of the second ratio, the step S06
Also include afterwards:The Logic Regression Models that training terminates are surveyed using the corresponding test set of Claims Resolution case type of selection
Examination, and test report is generated, for example, the test report generated after being tested includes test hit rate, coverage rate, accuracy rate etc.
Content.
Step S2, the incidence relation based on predetermined anti-fraud identification model and risks and assumptions obtains the Claims Resolution application pair
The risks and assumptions for the anti-fraud identification model answered, the corresponding anti-fraud of the Claims Resolution application is obtained from predetermined financial system and is known
The information of the risks and assumptions of other model;
In the present embodiment, risks and assumptions include but is not limited to throw the customer information of insurant, disease information, treatment letter
Breath, Claims Resolution case are in danger duration, insurer's credit information and overdue information.
In anti-fake system of settling a claim, each anti-fraud identification model, which has, associates at least one corresponding risks and assumptions,
Such as the first anti-fraud identification model corresponding with unexpected class Claims Resolution case of dieing, its risks and assumptions include throwing the visitor of insurant
Family information, Claims Resolution case are in danger duration, insurer's credit information and overdue information etc..Based on anti-fraud identification model and risk because
The incidence relation of son can determine the risks and assumptions of the corresponding anti-fraud identification model of Claims Resolution application.
Then the letter of the risks and assumptions of the corresponding anti-fraud identification model of Claims Resolution application is obtained from predetermined financial system
Breath, predetermined financial system includes but is not limited to life insurance and insured the outer of intra-company's systems such as system, Claims Resolution system and docking
Portion's information system;The information of risks and assumptions is, for example, the customer information for throwing insurant (such as including name, age, annual income
Information etc.), disease and treatment (such as the expense for suffering from which kind of disease, the duration treated and treatment), Claims Resolution case
Part is in danger duration (being, for example, 1 month), insurer's loan profile (such as including without loan, loan date and the amount of money) with exceeding
Phase information (for example, provide a loan and do not refund overdue 5 days) etc..
Step S3, the Claims Resolution is applied for the information of the risks and assumptions of corresponding anti-fraud identification model is inputted to this and counter taken advantage of
Cheat in identification model, corresponding risk indicator value is applied for export the Claims Resolution by the anti-fraud identification model;
Step S4, the corresponding wind of the Claims Resolution application is determined based on the risk indicator value and default risk processing rule
Dangerous processing mode, and risk processing is carried out to the Claims Resolution application based on the risk processing mode.
In the present embodiment, after the information of risks and assumptions of the corresponding anti-fraud identification model of Claims Resolution application is obtained, by institute
The information of the risks and assumptions of acquisition is inputted into the corresponding anti-fraud identification model of the Claims Resolution application, then exports the Claims Resolution application
Corresponding risk indicator value, risk indicator value is used to evaluate the degree that the client for carrying out Claims Resolution application has fraud.
Then, the corresponding risk processing mode of Claims Resolution application is determined based on risk indicator value, with the Claims Resolution application to client
Handled, if such as risk indicator value is higher, manual examination and verification, or if wind can be further carried out to Claims Resolution application
Dangerous desired value is relatively low, then can not carry out any processing to Claims Resolution application.
Compared with prior art, because the anti-fraud identification model of the present embodiment is to be trained life using big data in advance
Into, therefore, it is possible to recognize implicit risk of fraud to the full extent, strengthen the dynamics of implicit risk identification;In addition, using
Anti- fraud identification model obtains risk indicator value, to handle Claims Resolution application, can rapidly carry out the Claims Resolution Shen of batch
Processing please, improves the efficiency of anti-fraud identification.
In a preferred embodiment, on the basis of above-mentioned Fig. 1 embodiment, above-mentioned default risk processing rule
For:If the Claims Resolution applies for that corresponding risk indicator value is more than default risk indicator threshold value, risk indicator value is more than pre-
If the Claims Resolution of risk indicator threshold value apply for the case of artificial anti-fraud identifying processing;If corresponding wind is applied in the Claims Resolution
Dangerous desired value is less than or equal to default risk indicator threshold value and more than or equal to default secure threshold, then to the Claims Resolution of predetermined number
Application is inspected by random samples in default sampling observation ratio, and the case of artificial anti-fraud identifying processing is applied for the Claims Resolution inspected by random samples.
In the present embodiment, default risk indicator threshold value is more than default secure threshold, such as default risk indicator threshold
It is worth for 6, default secure threshold is 3.
If the corresponding risk indicator value of Claims Resolution application is more than default risk indicator threshold value, such as Claims Resolution application correspondence
Risk indicator value be 8, then the risk indicator value is the case that artificial anti-fraud identifying processing is applied in 8 corresponding Claims Resolutions;
If the corresponding risk indicator value of Claims Resolution application is less than or equal to default risk indicator threshold value and more than or equal to default safety
Threshold value, such as corresponding risk indicator value of Claims Resolution application is 5, then the corresponding risk indicator value of the Claims Resolution application is less than or equal to default
Risk indicator threshold value and more than or equal to default secure threshold, then the risk indicator value is that 5 corresponding Claims Resolution applications are added extremely
In case to be inspected by random samples, then inspected by random samples again in case to be inspected by random samples according to default sampling observation ratio, for example, waiting to inspect by random samples
Case in inspected by random samples according to 50% ratio, the case of artificial anti-fraud identifying processing is applied in the Claims Resolution inspected by random samples;
In addition, being less than the Claims Resolution application of default secure threshold for risk indicator value, then any processing can not be carried out to it.
Default risk processing rule can also be in addition:To each Claims Resolution application according to corresponding risk indicator value from greatly to
Small order is ranked up;Determine to preset corresponding first caseload of risk ratio (for example, 10 in all Claims Resolution applications
5) the corresponding Claims Resolution application caseload of 50% risk ratio is in Claims Resolution application;Determine preset security in all Claims Resolution applications
Corresponding second caseload of ratio is (for example, the corresponding Claims Resolution application case of 10% safe ratio in 10 Claims Resolution applications
1) quantity be;The case of artificial anti-fraud identifying processing is applied for the Claims Resolution of the first maximum caseload of risk indicator value
Part, the Claims Resolution application to the second minimum caseload of risk indicator value is not dealt with, and remaining Claims Resolution application is inspected by random samples,
Then artificial anti-fraud identifying processing is carried out to the Claims Resolution application inspected by random samples.
The present embodiment carries out artificial anti-fraud identifying processing for the maximum part Claims Resolution application of risk indicator value, for
The Claims Resolution application of risk indicator value larger portion is inspected by random samples, and artificial anti-fraud is then carried out to the Claims Resolution application inspected by random samples knows
Other places are managed, for the minimum part Claims Resolution application of risk indicator value without any processing, so to different risk indicator values
Claims Resolution application use different processing modes, it is possible to increase the efficiency of anti-fraud identification.
As shown in figure 3, structural representations of the Fig. 3 for the anti-embodiment of fraud identifying device one of settlement of insurance claim of the present invention, should
The anti-fraud identifying device of settlement of insurance claim includes:
First determining module 101, for after the Claims Resolution application for carrying Claims Resolution information is received, determining the Claims Resolution application
Corresponding Claims Resolution case type, the reason is obtained according to predetermined Claims Resolution case type and the incidence relation of anti-fraud identification model
Pay for the anti-fraud identification model of the corresponding training in advance generation of application;
In the present embodiment, the anti-fake system of settling a claim is installed in the anti-fraud identifying device of settlement of insurance claim, and client can be
Operated in the anti-fraud identifying device of settlement of insurance claim, log in the anti-fake system of Claims Resolution, to initiate Claims Resolution application, insurance reason
The anti-fraud identifying device paid for can be any suitable equipment such as mobile phone or computer.
Anti- fake system settle a claim after the Claims Resolution application of client is received, according to the Claims Resolution information carried in Claims Resolution application
(such as case type information of settling a claim, the customer information of throwing warrantee) determines the corresponding Claims Resolution case type of the Claims Resolution application,
Claims Resolution case type includes but is not limited to unexpected die class Claims Resolution case, disease and die class Claims Resolution case, major disease class reason
Pay for case and unexpected medical class Claims Resolution case.
In the present embodiment, training in advance generates multiple anti-fraud identification models and preserved in anti-fake system of settling a claim, counter to take advantage of
Cheating identification model is, for example,:The first anti-fraud identification model corresponding with unexpected class Claims Resolution case of dieing, class reason of being die with disease
Pay for the corresponding second anti-fraud identification model of case, identification model is instead cheated with major disease class Claims Resolution case the corresponding 3rd with
And instead cheat identification model with unexpected medical class Claims Resolution case the corresponding 4th.Each Claims Resolution case type association correspondence one is counter to take advantage of
Preserved in swindleness identification model, anti-fake system of settling a claim and pay for case type pass corresponding with the association one by one of anti-fraud identification model
System, it is determined that after the corresponding Claims Resolution case type of the Claims Resolution application, the corresponding training in advance generation of Claims Resolution application can be obtained
Anti- fraud identification model.
Preferably, also include with reference to the processing unit refering to Fig. 4, settlement of insurance claim:
Second determining module 011, for according to predetermined Claims Resolution case type and the mapping relations of the full dose factor, it is determined that choosing
Variable of the corresponding full dose factor of case type as sample data set of settling a claim selected;The full dose factor be prepare for modeling with
All data factors of the Claims Resolution case type correlation of correspondence Claims Resolution application, risks and assumptions are the subset of the full dose factor.
Demarcating module 012, sample data (such as 100,000 parts samples for preparing predetermined number for each Claims Resolution case type
Data), it is that each sample data demarcates corresponding risk indicator value, forms sample data set;
Extraction module 013, it is identified each for being extracted from the corresponding sample data of Claims Resolution case type of selection
The information of the full dose factor;The client for throwing insurant is for example extracted from the corresponding sample data of Claims Resolution case type of selection
Which kind of disease information (such as including name, age, annual income information), disease and treatment (for example suffer from, are treated
Duration and treatment expense etc.), be in danger duration (being, for example, 1 month), insurer's loan profile of Claims Resolution case (for example include
Without loan, loan date and the amount of money etc.) and overdue information (for example, provide a loan and do not refund overdue 5 days) etc..
Division module 014, for the information of the full dose factor of the corresponding sample data of Claims Resolution case type of selection to be divided
For the training set of the first ratio, the checking collection of the second ratio;For example for the corresponding sample data of each Claims Resolution case type, with
The sample data of 50% ratio collects as training set using the sample data of 25% ratio as checking.
Training module 015, corresponding logistic regression is trained for the corresponding training set of Claims Resolution case type using selection
Model, draws the incidence relation of the corresponding Logic Regression Models of Claims Resolution case type and risks and assumptions;Wherein, trained first
When, the parameter of Logic Regression Models is the parameter of acquiescence, and with the progress of training, parameter is constantly adjusted;Risks and assumptions are
The full dose factor of the specific part of corresponding Logic Regression Models selection.In addition, in the training process, different full dose factor pairs
Risk indicator value has different degrees of influence, can then will using some larger full dose factors of influence degree as risks and assumptions
These risks and assumptions are associated with Logic Regression Models.
Authentication module 016, for the corresponding checking collection checking of the Claims Resolution case type using the selection logistic regression mould
Hit rate, coverage rate, the accuracy rate of type, if the hit rate, coverage rate and accuracy rate are corresponded to respectively is more than or equal to default life
Middle rate, coverage rate and accuracy rate, default hit rate, coverage rate and accuracy rate are respectively for example 0.98,0.99 and 0.98, then instruct
White silk terminates;If the hit rate of the gained of verifying logic regression model is less than default hit rate, or verifying logic returns mould
The coverage rate of the gained of type be less than default coverage rate, or verifying logic regression model gained accuracy rate be less than it is default
Accuracy rate, then need to increase the quantity of the sample data of training set, and is trained again, until after training terminates, being tested
Card gained hit rate, coverage rate and accuracy rate correspond to be more than or equal to default hit rate, coverage rate and accuracy rate respectively, to train
The Logic Regression Models obtained after end are used as anti-fraud identification model to be used.
In addition, anti-fraud identification model is preferably that (Logic Regression Models are in identification Claims Resolution application for Logic Regression Models
Stronger, application scenarios are wider in terms of the interpretation of risk of fraud), certainly anti-fraud identification model can also be other moulds
Type, such as neural network model, decision-tree model, excessive restriction is not done herein.
Preferably, division module 014 is further used for the full dose of the corresponding sample data of Claims Resolution case type of selection
The information of the factor is divided into the training set of the first ratio, the checking collection and the test set of the 3rd ratio of the second ratio;The insurance reason
The processing unit of compensation also includes:Test module, the Logic Regression Models for terminating to training use the Claims Resolution case class of selection
The corresponding test set of type is tested, and generates test report, is ordered for example, the test report generated after being tested includes test
The contents such as middle rate, coverage rate, accuracy rate.
Acquisition module 102, obtains described for the incidence relation based on predetermined anti-fraud identification model and risks and assumptions
The risks and assumptions of the corresponding anti-fraud identification model of Claims Resolution application, obtain the Claims Resolution application correspondence from predetermined financial system
Anti- fraud identification model risks and assumptions information;
In the present embodiment, risks and assumptions include but is not limited to throw the customer information of insurant, disease information, treatment letter
Breath, Claims Resolution case are in danger duration, insurer's credit information and overdue information.
In anti-fake system of settling a claim, each anti-fraud identification model, which has, associates at least one corresponding risks and assumptions,
Such as the first anti-fraud identification model corresponding with unexpected class Claims Resolution case of dieing, its risks and assumptions include throwing the visitor of insurant
Family information, Claims Resolution case are in danger duration, insurer's credit information and overdue information etc..Based on anti-fraud identification model and risk because
The incidence relation of son can determine the risks and assumptions of the corresponding anti-fraud identification model of Claims Resolution application.
Then the letter of the risks and assumptions of the corresponding anti-fraud identification model of Claims Resolution application is obtained from predetermined financial system
Breath, predetermined financial system includes but is not limited to life insurance and insured the outer of intra-company's systems such as system, Claims Resolution system and docking
Portion's information system;The information of risks and assumptions is, for example, the customer information for throwing insurant (such as including name, age, annual income
Information etc.), disease and treatment (such as the expense for suffering from which kind of disease, the duration treated and treatment), Claims Resolution case
Part is in danger duration (being, for example, 1 month), insurer's loan profile (such as including without loan, loan date and the amount of money) with exceeding
Phase information (for example, provide a loan and do not refund overdue 5 days) etc..
Output module 103, for the Claims Resolution to be applied to, the information of risks and assumptions of corresponding anti-fraud identification model is defeated
Enter into the anti-fraud identification model, corresponding risk indicator is applied for export the Claims Resolution by the anti-fraud identification model
Value;
Processing module 104, for determining the Claims Resolution Shen based on the risk indicator value and default risk processing rule
Please corresponding risk processing mode, and based on the risk processing mode to it is described Claims Resolution application carry out risk processing.
In the present embodiment, after the information of risks and assumptions of the corresponding anti-fraud identification model of Claims Resolution application is obtained, by institute
The information of the risks and assumptions of acquisition is inputted into the corresponding anti-fraud identification model of the Claims Resolution application, then exports the Claims Resolution application
Corresponding risk indicator value, risk indicator value is used to evaluate the degree that the client for carrying out Claims Resolution application has fraud.
Then, the corresponding risk processing mode of Claims Resolution application is determined based on risk indicator value, with the Claims Resolution application to client
Handled, if such as risk indicator value is higher, manual examination and verification, or if wind can be further carried out to Claims Resolution application
Dangerous desired value is relatively low, then can not carry out any processing to Claims Resolution application.
Compared with prior art, because the anti-fraud identification model of the present embodiment is to be trained life using big data in advance
Into, therefore, it is possible to recognize implicit risk of fraud to the full extent;Refer in addition, obtaining risk using anti-fraud identification model
Scale value can rapidly carry out the processing of the Claims Resolution application of batch to handle Claims Resolution application, improve anti-fraud identification
Efficiency.
In a preferred embodiment, on the basis of above-mentioned Fig. 3 embodiment, the default risk processing rule is:
If the Claims Resolution applies for that corresponding risk indicator value is more than default risk indicator threshold value, risk indicator value is more than default
The case of artificial anti-fraud identifying processing is applied in the Claims Resolution of risk indicator threshold value;If the Claims Resolution applies for that corresponding risk refers to
Scale value is less than or equal to default risk indicator threshold value and more than or equal to default secure threshold, then to the Claims Resolution application of predetermined number
Inspected by random samples in default sampling observation ratio, the case of artificial anti-fraud identifying processing is applied for the Claims Resolution inspected by random samples.
In the present embodiment, default risk indicator threshold value is more than default secure threshold, such as default risk indicator threshold
It is worth for 6, default secure threshold is 3.
If the corresponding risk indicator value of Claims Resolution application is more than default risk indicator threshold value, such as Claims Resolution application correspondence
Risk indicator value be 8, then the risk indicator value is the case that artificial anti-fraud identifying processing is applied in 8 corresponding Claims Resolutions;
If the corresponding risk indicator value of Claims Resolution application is less than or equal to default risk indicator threshold value and more than or equal to default safety
Threshold value, such as corresponding risk indicator value of Claims Resolution application is 5, then the corresponding risk indicator value of the Claims Resolution application is less than or equal to default
Risk indicator threshold value and more than or equal to default secure threshold, then the risk indicator value is that 5 corresponding Claims Resolution applications are added extremely
In case to be inspected by random samples, then inspected by random samples again in case to be inspected by random samples according to default sampling observation ratio, for example, waiting to inspect by random samples
Case in inspected by random samples according to 50% ratio, the case of artificial anti-fraud identifying processing is applied in the Claims Resolution inspected by random samples;
In addition, being less than the Claims Resolution application of default secure threshold for risk indicator value, then any processing can not be carried out to it.
Default risk processing rule can also be in addition:To each Claims Resolution application according to corresponding risk indicator value from greatly to
Small order is ranked up;Determine to preset corresponding first caseload of risk ratio (for example, 10 in all Claims Resolution applications
5) the corresponding Claims Resolution application caseload of 50% risk ratio is in Claims Resolution application;Determine preset security in all Claims Resolution applications
Corresponding second caseload of ratio is (for example, the corresponding Claims Resolution application case of 10% safe ratio in 10 Claims Resolution applications
1) quantity be;The case of artificial anti-fraud identifying processing is applied for the Claims Resolution of the first maximum caseload of risk indicator value
Part, the Claims Resolution application to the second minimum caseload of risk indicator value is not dealt with, and remaining Claims Resolution application is inspected by random samples,
Then artificial anti-fraud identifying processing is carried out to the Claims Resolution application inspected by random samples.
The present embodiment carries out artificial anti-fraud identifying processing for the maximum part Claims Resolution application of risk indicator value, for
The Claims Resolution application of risk indicator value larger portion is inspected by random samples, and artificial anti-fraud is then carried out to the Claims Resolution application inspected by random samples knows
Other places are managed, for the minimum part Claims Resolution application of risk indicator value without any processing, so to different risk indicator values
Claims Resolution application use different processing modes, it is possible to increase the efficiency of anti-fraud identification.
The foregoing is only presently preferred embodiments of the present invention, be not intended to limit the invention, it is all the present invention spirit and
Within principle, any modification, equivalent substitution and improvements made etc. should be included in the scope of the protection.
Claims (10)
1. a kind of anti-fraud recognition methods of settlement of insurance claim, it is characterised in that the anti-fraud recognition methods bag of the settlement of insurance claim
Include:
S1, settles a claim anti-fake system after the Claims Resolution application for carrying Claims Resolution information is received, and determines that the Claims Resolution application is corresponding
Claims Resolution case type, applies according to predetermined Claims Resolution case type and the incidence relation acquisition Claims Resolution of anti-fraud identification model
The anti-fraud identification model of corresponding training in advance generation;
S2, the incidence relation based on predetermined anti-fraud identification model and risks and assumptions obtains that the Claims Resolution application is corresponding counter to take advantage of
The risks and assumptions of identification model are cheated, the corresponding anti-fraud identification model of the Claims Resolution application is obtained from predetermined financial system
The information of risks and assumptions;
S3, the Claims Resolution is applied for the information of the risks and assumptions of corresponding anti-fraud identification model is inputted to the anti-fraud identification mould
In type, corresponding risk indicator value is applied for export the Claims Resolution by the anti-fraud identification model;
S4, the corresponding risk processing side of the Claims Resolution application is determined based on the risk indicator value and default risk processing rule
Formula, and risk processing is carried out to the Claims Resolution application based on the risk processing mode.
2. the anti-fraud recognition methods of settlement of insurance claim according to claim 1, it is characterised in that the Claims Resolution case type
Die class Claims Resolution case, major disease class Claims Resolution case and unexpected medical class Claims Resolution including unexpected die class Claims Resolution case, disease
Case.
3. the anti-fraud recognition methods of settlement of insurance claim according to claim 1, it is characterised in that the default risk processing
Rule is:
If the Claims Resolution applies for that corresponding risk indicator value is more than default risk indicator threshold value, risk indicator value is more than pre-
If the Claims Resolution of risk indicator threshold value apply for the case of artificial anti-fraud identifying processing;
If the Claims Resolution applies for that corresponding risk indicator value is less than or equal to default risk indicator threshold value and more than or equal to default
Secure threshold, the then Claims Resolution application to predetermined number is inspected by random samples in default sampling observation ratio, is made with the Claims Resolution application inspected by random samples
For the case of artificial anti-fraud identifying processing.
4. the anti-fraud recognition methods of the settlement of insurance claim according to any one of claims 1 to 3, it is characterised in that described anti-
Fraud identification model includes before being Logic Regression Models, the step S1:
S01, according to predetermined Claims Resolution case type and the mapping relations of the full dose factor, determines the Claims Resolution case type correspondence of selection
The full dose factor as sample data set variable;
S02, is the sample data that each Claims Resolution case type prepares predetermined number, is that the corresponding risk of each sample data demarcation refers to
Scale value, forms sample data set;
S03, extracts the information of identified each full dose factor from the corresponding sample data of Claims Resolution case type of selection;
S04, the information of the full dose factor of the corresponding sample data of Claims Resolution case type of selection is divided into the training of the first ratio
Collection, the checking collection of the second ratio;
S05, trains corresponding Logic Regression Models using the corresponding training set of the Claims Resolution case type of selection, draws the Claims Resolution case
The incidence relation of the corresponding Logic Regression Models of part type and risks and assumptions;
S06, the hit rate of the Logic Regression Models, covering are verified using the corresponding checking collection of the Claims Resolution case type of selection
Rate, accuracy rate, if the hit rate, coverage rate and accuracy rate are corresponded to respectively is more than or equal to default hit rate, coverage rate and standard
True rate, then training terminates, otherwise, increases the quantity of the sample data of training set and is trained again.
5. the anti-fraud recognition methods of settlement of insurance claim according to claim 4, it is characterised in that the step S04 enters one
Step includes:The information of the full dose factor of the corresponding sample data of Claims Resolution case type of selection is divided into the training of the first ratio
Also include after the checking collection and the test set of the 3rd ratio of collection, the second ratio, the step S06:
The Logic Regression Models that training terminates are tested using the corresponding test set of Claims Resolution case type of selection, and generated
Test report.
6. a kind of anti-fraud identifying device of settlement of insurance claim, it is characterised in that the anti-fraud identifying device bag of the settlement of insurance claim
Include:
First determining module, for after the Claims Resolution application for carrying Claims Resolution information is received, determining that the Claims Resolution application is corresponding
Claims Resolution case type, applies according to predetermined Claims Resolution case type and the incidence relation acquisition Claims Resolution of anti-fraud identification model
The anti-fraud identification model of corresponding training in advance generation;
Acquisition module, the Claims Resolution application is obtained for the incidence relation based on predetermined anti-fraud identification model and risks and assumptions
The risks and assumptions of corresponding anti-fraud identification model, obtain the corresponding anti-fraud of the Claims Resolution application from predetermined financial system
The information of the risks and assumptions of identification model;
Output module, for the Claims Resolution to be applied to, it is anti-to this that the information of risks and assumptions of corresponding anti-fraud identification model is inputted
Cheat in identification model, corresponding risk indicator value is applied for export the Claims Resolution by the anti-fraud identification model;
Processing module, for determining that the Claims Resolution application is corresponding based on the risk indicator value and default risk processing rule
Risk processing mode, and risk processing is carried out to the Claims Resolution application based on the risk processing mode.
7. the processing unit of settlement of insurance claim according to claim 6, it is characterised in that the Claims Resolution case type includes meaning
Outer die class Claims Resolution case, disease are die class Claims Resolution case, major disease class Claims Resolution case and unexpected medical class Claims Resolution case.
8. the processing unit of settlement of insurance claim according to claim 6, it is characterised in that the default risk processing rule
For:
If the Claims Resolution applies for that corresponding risk indicator value is more than default risk indicator threshold value, risk indicator value is more than pre-
If the Claims Resolution of risk indicator threshold value apply for the case of artificial anti-fraud identifying processing;If corresponding wind is applied in the Claims Resolution
Dangerous desired value is less than or equal to default risk indicator threshold value and more than or equal to default secure threshold, then to the Claims Resolution of predetermined number
Application is inspected by random samples in default sampling observation ratio, and the case of artificial anti-fraud identifying processing is applied for the Claims Resolution inspected by random samples.
9. the processing unit of the settlement of insurance claim according to any one of claim 6 to 8, it is characterised in that the settlement of insurance claim
Processing unit also include:
Second determining module, for according to predetermined Claims Resolution case type and the mapping relations of the full dose factor, determining the reason of selection
The corresponding full dose factor of case type is paid for as the variable of sample data set;
Demarcating module, the sample data for preparing predetermined number for each Claims Resolution case type is each sample data demarcation pair
The risk indicator value answered, forms sample data set;
Extraction module, for extracting identified each full dose factor from the corresponding sample data of Claims Resolution case type of selection
Information;
Division module, for the information of the full dose factor of the corresponding sample data of Claims Resolution case type of selection to be divided into the first ratio
Training set, the checking collection of the second ratio of example;
Training module, trains corresponding Logic Regression Models for the corresponding training set of Claims Resolution case type using selection, obtains
Go out the incidence relation of the corresponding Logic Regression Models of Claims Resolution case type and risks and assumptions;
Authentication module, the hit for the Claims Resolution case type corresponding checking collection checking Logic Regression Models using selection
Rate, coverage rate, accuracy rate, if the hit rate, coverage rate and accuracy rate are corresponded to respectively is more than or equal to default hit rate, covering
Rate and accuracy rate, then training terminate, otherwise, increase the quantity of the sample data of training set and are trained again.
10. the processing unit of settlement of insurance claim according to claim 9, it is characterised in that the division module is further used
In the information of the full dose factor of the corresponding sample data of Claims Resolution case type of selection is divided into the training set of the first ratio, second
The checking collection and the test set of the 3rd ratio of ratio;
The processing unit of the settlement of insurance claim also includes:Test module, the Logic Regression Models for terminating to training use choosing
The corresponding test set of Claims Resolution case type selected is tested, and generates test report.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710365366.2A CN107240024A (en) | 2017-05-22 | 2017-05-22 | The anti-fraud recognition methods of settlement of insurance claim and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710365366.2A CN107240024A (en) | 2017-05-22 | 2017-05-22 | The anti-fraud recognition methods of settlement of insurance claim and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107240024A true CN107240024A (en) | 2017-10-10 |
Family
ID=59985548
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710365366.2A Pending CN107240024A (en) | 2017-05-22 | 2017-05-22 | The anti-fraud recognition methods of settlement of insurance claim and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107240024A (en) |
Cited By (33)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107871285A (en) * | 2017-12-06 | 2018-04-03 | 和金在线(北京)科技有限公司 | A kind of health insurance pays for the method for detecting and system of fraud and abuse |
CN107909299A (en) * | 2017-12-11 | 2018-04-13 | 凯泰铭科技(北京)有限公司 | People hinders Claims Resolution data risk checking method and system |
CN108170909A (en) * | 2017-12-13 | 2018-06-15 | 中国平安财产保险股份有限公司 | Model output method, equipment and the storage medium of a kind of intelligent modeling |
CN108182515A (en) * | 2017-12-13 | 2018-06-19 | 中国平安财产保险股份有限公司 | Intelligent rules engine rule output method, equipment and computer readable storage medium |
CN108346036A (en) * | 2018-03-06 | 2018-07-31 | 平安科技(深圳)有限公司 | Insurance policy concentrates vouching method, electronic device and readable storage medium storing program for executing |
CN108492196A (en) * | 2018-03-08 | 2018-09-04 | 平安医疗健康管理股份有限公司 | The air control method of medical insurance unlawful practice is inferred by data analysis |
CN108694521A (en) * | 2018-07-02 | 2018-10-23 | 平安健康保险股份有限公司 | Individual risk's pricing method and system |
CN108694520A (en) * | 2018-07-02 | 2018-10-23 | 平安健康保险股份有限公司 | Claims Resolution customer risk recognition methods and system |
CN108734591A (en) * | 2018-04-09 | 2018-11-02 | 中国平安人寿保险股份有限公司 | Cheat appraisal procedure, device, storage medium and the terminal of case |
CN108876636A (en) * | 2018-06-19 | 2018-11-23 | 平安健康保险股份有限公司 | The intelligent air control method of Claims Resolution, system, computer equipment and storage medium |
CN109191312A (en) * | 2018-08-07 | 2019-01-11 | 阳光财产保险股份有限公司 | A kind of anti-fraud air control method and device of Claims Resolution |
CN109242107A (en) * | 2018-09-11 | 2019-01-18 | 北京芯盾时代科技有限公司 | Anti- fraud model training method, system based on transfer learning and counter cheat method |
CN109272241A (en) * | 2018-09-29 | 2019-01-25 | 北京优全智汇信息技术有限公司 | One seed nucleus pays for method and apparatus |
CN109409648A (en) * | 2018-09-10 | 2019-03-01 | 平安科技(深圳)有限公司 | Claims Resolution air control method, apparatus, computer equipment and storage medium |
CN109413031A (en) * | 2018-08-31 | 2019-03-01 | 深圳壹账通智能科技有限公司 | Construction method, device, equipment and the readable storage medium storing program for executing of anti-fraud model |
CN109461068A (en) * | 2018-09-13 | 2019-03-12 | 深圳壹账通智能科技有限公司 | Judgment method, device, equipment and the computer readable storage medium of fraud |
CN109543925A (en) * | 2019-01-07 | 2019-03-29 | 平安科技(深圳)有限公司 | Risk Forecast Method, device, computer equipment and storage medium based on machine learning |
CN109636305A (en) * | 2018-10-29 | 2019-04-16 | 平安医疗健康管理股份有限公司 | Compensate method of adjustment, device, equipment and the readable storage medium storing program for executing of the amount of money |
CN109859059A (en) * | 2019-01-17 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Settlement of insurance claim method, apparatus, computer equipment and storage medium |
CN110009511A (en) * | 2019-01-29 | 2019-07-12 | 阿里巴巴集团控股有限公司 | Arbitrage behavior recognition methods, arbitrage behavior identification model training method and system |
CN110033385A (en) * | 2019-03-05 | 2019-07-19 | 阿里巴巴集团控股有限公司 | The method, apparatus and electronic equipment of information processing |
CN110136010A (en) * | 2019-04-18 | 2019-08-16 | 中国平安财产保险股份有限公司 | The method, apparatus and computer equipment of risk case are judged based on neural network |
CN110175928A (en) * | 2019-05-14 | 2019-08-27 | 安泰保险科技(深圳)有限公司 | A kind of insurance that the intentional manufacture vehicle of identification mutually touches insurance risk is counter to cheat method |
CN110689441A (en) * | 2019-08-15 | 2020-01-14 | 平安健康保险股份有限公司 | Case auditing method, device and equipment based on commonality extraction and storage medium |
CN110689443A (en) * | 2019-09-02 | 2020-01-14 | 中国平安财产保险股份有限公司 | Insurance data processing method and device, storage medium and server |
CN110717821A (en) * | 2019-09-09 | 2020-01-21 | 上海凯京信达科技集团有限公司 | Vehicle loan assessment method and device, computer storage medium and electronic equipment |
CN110909753A (en) * | 2018-09-14 | 2020-03-24 | 中国移动通信集团广东有限公司 | Data classification method, system and equipment |
CN111861762A (en) * | 2020-07-28 | 2020-10-30 | 贵州力创科技发展有限公司 | Data processing method and system for anti-fraud recognition of vehicle insurance |
CN112508711A (en) * | 2020-11-11 | 2021-03-16 | 中国人寿保险股份有限公司 | Automatic claim checking method and related equipment for policy claim settlement |
CN112950397A (en) * | 2021-05-17 | 2021-06-11 | 太平金融科技服务(上海)有限公司深圳分公司 | Claims risk estimation method and device, computer equipment and storage medium |
CN113469826A (en) * | 2021-07-22 | 2021-10-01 | 阳光人寿保险股份有限公司 | Information processing method, device, equipment and storage medium |
CN113487440A (en) * | 2021-07-26 | 2021-10-08 | 联仁健康医疗大数据科技股份有限公司 | Model generation method, health insurance claim settlement determination method, device, equipment and medium |
CN114723570A (en) * | 2021-01-05 | 2022-07-08 | 中国银行保险信息技术管理有限公司 | Agricultural risk monitoring method and device, storage medium and computer equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104679777A (en) * | 2013-12-02 | 2015-06-03 | 中国银联股份有限公司 | Method and system for detecting fraudulent trading |
CN105590261A (en) * | 2014-12-31 | 2016-05-18 | 中国银联股份有限公司 | Merchant risk estimation method and system |
CN106529960A (en) * | 2016-11-07 | 2017-03-22 | 中国银联股份有限公司 | Fraud transaction detection method for electronic transaction |
-
2017
- 2017-05-22 CN CN201710365366.2A patent/CN107240024A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104679777A (en) * | 2013-12-02 | 2015-06-03 | 中国银联股份有限公司 | Method and system for detecting fraudulent trading |
CN105590261A (en) * | 2014-12-31 | 2016-05-18 | 中国银联股份有限公司 | Merchant risk estimation method and system |
CN106529960A (en) * | 2016-11-07 | 2017-03-22 | 中国银联股份有限公司 | Fraud transaction detection method for electronic transaction |
Cited By (42)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107871285A (en) * | 2017-12-06 | 2018-04-03 | 和金在线(北京)科技有限公司 | A kind of health insurance pays for the method for detecting and system of fraud and abuse |
CN107909299A (en) * | 2017-12-11 | 2018-04-13 | 凯泰铭科技(北京)有限公司 | People hinders Claims Resolution data risk checking method and system |
CN108170909A (en) * | 2017-12-13 | 2018-06-15 | 中国平安财产保险股份有限公司 | Model output method, equipment and the storage medium of a kind of intelligent modeling |
CN108182515A (en) * | 2017-12-13 | 2018-06-19 | 中国平安财产保险股份有限公司 | Intelligent rules engine rule output method, equipment and computer readable storage medium |
CN108170909B (en) * | 2017-12-13 | 2021-08-03 | 中国平安财产保险股份有限公司 | Intelligent modeling model output method, equipment and storage medium |
CN108346036A (en) * | 2018-03-06 | 2018-07-31 | 平安科技(深圳)有限公司 | Insurance policy concentrates vouching method, electronic device and readable storage medium storing program for executing |
CN108492196A (en) * | 2018-03-08 | 2018-09-04 | 平安医疗健康管理股份有限公司 | The air control method of medical insurance unlawful practice is inferred by data analysis |
CN108734591A (en) * | 2018-04-09 | 2018-11-02 | 中国平安人寿保险股份有限公司 | Cheat appraisal procedure, device, storage medium and the terminal of case |
CN108876636A (en) * | 2018-06-19 | 2018-11-23 | 平安健康保险股份有限公司 | The intelligent air control method of Claims Resolution, system, computer equipment and storage medium |
CN108876636B (en) * | 2018-06-19 | 2023-10-27 | 平安健康保险股份有限公司 | Intelligent air control method, system, computer equipment and storage medium for claim settlement |
CN108694520A (en) * | 2018-07-02 | 2018-10-23 | 平安健康保险股份有限公司 | Claims Resolution customer risk recognition methods and system |
CN108694521A (en) * | 2018-07-02 | 2018-10-23 | 平安健康保险股份有限公司 | Individual risk's pricing method and system |
CN109191312A (en) * | 2018-08-07 | 2019-01-11 | 阳光财产保险股份有限公司 | A kind of anti-fraud air control method and device of Claims Resolution |
CN109413031A (en) * | 2018-08-31 | 2019-03-01 | 深圳壹账通智能科技有限公司 | Construction method, device, equipment and the readable storage medium storing program for executing of anti-fraud model |
CN109409648A (en) * | 2018-09-10 | 2019-03-01 | 平安科技(深圳)有限公司 | Claims Resolution air control method, apparatus, computer equipment and storage medium |
CN109242107A (en) * | 2018-09-11 | 2019-01-18 | 北京芯盾时代科技有限公司 | Anti- fraud model training method, system based on transfer learning and counter cheat method |
CN109461068A (en) * | 2018-09-13 | 2019-03-12 | 深圳壹账通智能科技有限公司 | Judgment method, device, equipment and the computer readable storage medium of fraud |
CN110909753A (en) * | 2018-09-14 | 2020-03-24 | 中国移动通信集团广东有限公司 | Data classification method, system and equipment |
CN110909753B (en) * | 2018-09-14 | 2022-07-01 | 中国移动通信集团广东有限公司 | Data classification method, system and equipment |
CN109272241A (en) * | 2018-09-29 | 2019-01-25 | 北京优全智汇信息技术有限公司 | One seed nucleus pays for method and apparatus |
CN109636305A (en) * | 2018-10-29 | 2019-04-16 | 平安医疗健康管理股份有限公司 | Compensate method of adjustment, device, equipment and the readable storage medium storing program for executing of the amount of money |
CN109543925B (en) * | 2019-01-07 | 2023-01-24 | 平安科技(深圳)有限公司 | Risk prediction method and device based on machine learning, computer equipment and storage medium |
CN109543925A (en) * | 2019-01-07 | 2019-03-29 | 平安科技(深圳)有限公司 | Risk Forecast Method, device, computer equipment and storage medium based on machine learning |
CN109859059A (en) * | 2019-01-17 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Settlement of insurance claim method, apparatus, computer equipment and storage medium |
CN110009511A (en) * | 2019-01-29 | 2019-07-12 | 阿里巴巴集团控股有限公司 | Arbitrage behavior recognition methods, arbitrage behavior identification model training method and system |
CN110033385A (en) * | 2019-03-05 | 2019-07-19 | 阿里巴巴集团控股有限公司 | The method, apparatus and electronic equipment of information processing |
CN110136010A (en) * | 2019-04-18 | 2019-08-16 | 中国平安财产保险股份有限公司 | The method, apparatus and computer equipment of risk case are judged based on neural network |
CN110175928A (en) * | 2019-05-14 | 2019-08-27 | 安泰保险科技(深圳)有限公司 | A kind of insurance that the intentional manufacture vehicle of identification mutually touches insurance risk is counter to cheat method |
CN110689441B (en) * | 2019-08-15 | 2023-06-27 | 平安健康保险股份有限公司 | Case auditing method, device, equipment and storage medium based on commonality extraction |
CN110689441A (en) * | 2019-08-15 | 2020-01-14 | 平安健康保险股份有限公司 | Case auditing method, device and equipment based on commonality extraction and storage medium |
CN110689443A (en) * | 2019-09-02 | 2020-01-14 | 中国平安财产保险股份有限公司 | Insurance data processing method and device, storage medium and server |
CN110717821A (en) * | 2019-09-09 | 2020-01-21 | 上海凯京信达科技集团有限公司 | Vehicle loan assessment method and device, computer storage medium and electronic equipment |
CN111861762B (en) * | 2020-07-28 | 2024-04-26 | 贵州力创科技发展有限公司 | Data processing method and system for identifying anti-fraud safety of vehicle |
CN111861762A (en) * | 2020-07-28 | 2020-10-30 | 贵州力创科技发展有限公司 | Data processing method and system for anti-fraud recognition of vehicle insurance |
CN112508711A (en) * | 2020-11-11 | 2021-03-16 | 中国人寿保险股份有限公司 | Automatic claim checking method and related equipment for policy claim settlement |
CN114723570A (en) * | 2021-01-05 | 2022-07-08 | 中国银行保险信息技术管理有限公司 | Agricultural risk monitoring method and device, storage medium and computer equipment |
CN114723570B (en) * | 2021-01-05 | 2024-02-09 | 中国银行保险信息技术管理有限公司 | Agricultural risk monitoring method and device, storage medium and computer equipment |
CN112950397A (en) * | 2021-05-17 | 2021-06-11 | 太平金融科技服务(上海)有限公司深圳分公司 | Claims risk estimation method and device, computer equipment and storage medium |
CN112950397B (en) * | 2021-05-17 | 2021-07-27 | 太平金融科技服务(上海)有限公司深圳分公司 | Claims risk estimation method and device, computer equipment and storage medium |
CN113469826B (en) * | 2021-07-22 | 2022-12-09 | 阳光人寿保险股份有限公司 | Information processing method, device, equipment and storage medium |
CN113469826A (en) * | 2021-07-22 | 2021-10-01 | 阳光人寿保险股份有限公司 | Information processing method, device, equipment and storage medium |
CN113487440A (en) * | 2021-07-26 | 2021-10-08 | 联仁健康医疗大数据科技股份有限公司 | Model generation method, health insurance claim settlement determination method, device, equipment and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107240024A (en) | The anti-fraud recognition methods of settlement of insurance claim and device | |
CN107657525A (en) | One kind loan measures and procedures for the examination and approval and server | |
CN109461070A (en) | A kind of risk measures and procedures for the examination and approval, device, storage medium and server | |
DE10145978A1 (en) | Personal authentication method for Internet based business transactions, involves collating received fingerprint data with stored reference fingerprint data based on which owner of card is judged | |
CN108038701A (en) | A kind of integrated study is counter to cheat test method and system | |
CN116342259A (en) | Automatic user credit rating method and device, electronic equipment and medium | |
CN105894380A (en) | Digital currency credit investigation system and method | |
CN109740869A (en) | Data checking method, device, computer equipment and storage medium | |
CN108416677A (en) | The method and device of Claims Resolution investigation | |
CN110851677A (en) | Reimbursement certificate processing method, device, terminal and computer readable storage medium | |
CN112199575A (en) | Virtual bank account opening method, device, equipment and computer storage medium | |
CN109598630A (en) | The monitoring and managing method and relevant device of fixed point medical institution based on living things feature recognition | |
CN112102073A (en) | Credit risk control method and system, electronic device and readable storage medium | |
CN110689443A (en) | Insurance data processing method and device, storage medium and server | |
CN110197427A (en) | Method and system are borrowed or lent money on a kind of line | |
CN106991312A (en) | Internet based on Application on Voiceprint Recognition is counter to cheat authentication method | |
CN111062834A (en) | Dispute case entity identification method and device, computer equipment and storage medium | |
CN111882140A (en) | Risk evaluation method, model training method, device, equipment and storage medium | |
CN111191889A (en) | Scoring card development method based on logistic regression and voting type model integration | |
DE102014100463A1 (en) | A method of identifying a user using a communication device | |
CN111680923A (en) | Occupational risk assessment method | |
CN114139931A (en) | Enterprise data evaluation method and device, computer equipment and storage medium | |
CN111160695A (en) | Method, system, device and storage medium for identifying risk account of computer operation | |
CN111753913A (en) | Counterfeit money classification method and device, electronic equipment and computer readable storage medium | |
CN115564591A (en) | Financing product determination method and related equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171010 |