CN109064343A

CN109064343A - Risk model method for building up, risk matching process, device, equipment and medium

Info

Publication number: CN109064343A
Application number: CN201810915907.9A
Authority: CN
Inventors: 丘健
Original assignee: Ping An Life Insurance Company of China Ltd
Current assignee: Ping An Life Insurance Company of China Ltd
Priority date: 2018-08-13
Filing date: 2018-08-13
Publication date: 2018-12-21
Anticipated expiration: 2038-08-13
Also published as: CN109064343B

Abstract

The invention discloses a kind of risk model method for building up, risk matching process, equipment and medium, which includes: to obtain at least one Claims Resolution information table, the corresponding Customer ID of each Claims Resolution information table and an insurance kind type；The corresponding first Claims Resolution frequency of each Customer ID and the corresponding second Claims Resolution frequency of each insurance kind type are counted, target customer ID and target insurance kind type are determined according to the first Claims Resolution frequency and the second Claims Resolution frequency；Information table of insuring is inquired according to target customer ID, obtains client characteristics dimension, and according to target insurance kind type queries insurance kind information table, obtains insurance kind characteristic dimension；It according to client characteristics dimension and insurance kind characteristic dimension, is modeled using decision Tree algorithms, obtains risk model.The risk model method for building up efficiently solves the problems, such as manually to verify speed slower and time-consuming.

Description

Risk model method for building up, risk matching process, device, equipment and medium

Technical field

The present invention relates to field of computer technology more particularly to a kind of risk model method for building up, risk matching process, dresses It sets, computer equipment and storage medium.

Background technique

Currently, more and more people select purchase insurance, for insurance company, customer insured is with the presence or absence of Claims Resolution wind Danger is mainly manually verified, and some verifies insufficient or subjective judgement fault by manually verifying to exist Situation, the time that not only labor intensive is verified, and it is slower by manually verifying speed.

Summary of the invention

According to this, it is necessary in view of the above technical problems, provide a kind of risk model that can quickly establish risk model Method for building up, device, computer equipment and storage medium.

A kind of risk model method for building up, comprising:

At least one Claims Resolution information table is obtained, each Claims Resolution information table corresponds to a Customer ID and an insurance kind type；

Count the corresponding first Claims Resolution frequency of each Customer ID and corresponding second Claims Resolution of each insurance kind type The frequency determines target customer ID and target insurance kind type according to the first Claims Resolution frequency and the second Claims Resolution frequency；

Information table of insuring is inquired according to the target customer ID, obtains client characteristics dimension, and according to the target insurance kind Type queries insurance kind information table obtains insurance kind characteristic dimension；

It according to the client characteristics dimension and the insurance kind characteristic dimension, is modeled using decision Tree algorithms, obtains wind Dangerous model.

A kind of risk model establishes device, comprising:

Data obtaining module, for obtaining at least one Claims Resolution information table, each corresponding client of the Claims Resolution information table ID and an insurance kind type；

Target information determining module, for counting the corresponding first Claims Resolution frequency of each Customer ID and each danger The corresponding second Claims Resolution frequency of seed type determines target customer ID according to the first Claims Resolution frequency and the second Claims Resolution frequency With target insurance kind type；

Characteristic dimension obtains module, for inquiring information table of insuring according to the target customer ID, obtains client characteristics dimension Degree, and according to the target insurance kind type queries insurance kind information table, obtain insurance kind characteristic dimension；

Risk model establishes module, is used for according to the client characteristics dimension and the insurance kind characteristic dimension, using decision Tree algorithm is modeled, and risk model is obtained.

A kind of computer equipment, including memory, processor and storage can be run in memory and on a processor Computer program, the step of realizing above-mentioned risk model method for building up when processor executes computer program.

A kind of computer readable storage medium, computer-readable recording medium storage have computer program, computer program The step of above-mentioned risk model method for building up is realized when being executed by processor.

According to this, it is necessary in view of the above technical problems, provide a kind of risk matching that can quickly carry out risk assessment Method, apparatus, computer equipment and storage medium.

A kind of risk matching process, comprising:

Obtain information to be matched of insuring；

Feature extraction is carried out to the information to be matched of insuring, obtains target customer's characteristic value and target insurance kind feature respectively Value；

The risk model obtained using the risk model method for building up, to target customer's characteristic value and the target Insurance kind characteristic value carries out risk matching, obtains risk assessment value；

If the risk assessment value is greater than default value-at-risk, the information to be matched of insuring is high risk information.

A kind of risk coalignment, comprising:

Data obtaining module, for obtaining information to be matched of insuring；

Characteristics extraction module obtains target customer for carrying out feature extraction to the information to be matched of insuring respectively Characteristic dimension and target insurance kind dimension；

Risk matching module, it is special to the target customer for the risk model that the risk model method for building up obtains It levies dimension and the target insurance kind dimension carries out risk matching, obtain risk assessment value；

High risk information object information determination module, it is described if being greater than default value-at-risk for the risk assessment value Information to be matched of insuring is high risk information.

A kind of computer equipment, including memory, processor and storage can be run in memory and on a processor Computer program, the step of realizing above-mentioned risk matching process when processor executes computer program.

A kind of computer readable storage medium, computer-readable recording medium storage have computer program, and processor executes The step of above-mentioned risk matching process is realized when computer program.

Above-mentioned risk model creation method, device, computer equipment and storage medium, by obtaining information table of settling a claim, with Just the Claims Resolution frequency of Customer ID and insurance kind type in each Claims Resolution information table is counted, and according to the Claims Resolution of Customer ID and insurance kind type The frequency determines target customer ID and target insurance kind type；Client characteristics dimension is determined according to target customer ID and target insurance kind type With insurance kind characteristic dimension, so as to the client characteristics dimension and insurance kind characteristic dimension of extraction be occur Claims Resolution the higher feature dimensions of the frequency Degree, models client characteristics dimension and insurance kind characteristic dimension by decision Tree algorithms, can be quickly obtained risk model, so as to It is subsequent to verify whether information of insuring is high risk information according to the risk model.

Above-mentioned risk matching process, device, computer equipment and storage medium, by obtaining information to be matched of insuring, with It will pass through risk model and risk assessment carried out to information to be matched of insuring；Feature extraction is carried out to information to be matched of insuring, respectively Target customer's characteristic value and target insurance kind characteristic value are obtained, to be matched subsequently through the characteristic value of extraction；Using risk Model carries out risk matching, energy quick obtaining risk assessment value, according to wind to target customer's characteristic value and target insurance kind characteristic value Dangerous assessed value determines that band matches whether information of insuring is high risk information, to assist insurance company quick lock in high risk user, To take timely measure.

Detailed description of the invention

In order to illustrate the technical solution of the embodiments of the present invention more clearly, below by institute in the description to the embodiment of the present invention Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention Example, for those of ordinary skill in the art, without any creative labor, can also be according to these attached drawings Obtain other attached drawings.

Fig. 1 is the application environment signal of one embodiment of the invention risk method for establishing model or risk matching process Figure；

Fig. 2 is a flow chart of one embodiment of the invention risk method for establishing model；

Fig. 3 is a flow chart of one embodiment of the invention risk method for establishing model；

Fig. 4 is a flow chart of one embodiment of the invention risk method for establishing model；

Fig. 5 is a flow chart of one embodiment of the invention risk method for establishing model；

Fig. 6 is a flow chart of one embodiment of the invention risk method for establishing model；

Fig. 7 is a flow chart of one embodiment of the invention risk matching process；

Fig. 8 is a schematic diagram of one embodiment of the invention risk model foundation device；

Fig. 9 is a schematic diagram of computer equipment in one embodiment of the invention；

Figure 10 is a schematic diagram of one embodiment of the invention risk coalignment.

Specific embodiment

Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete Site preparation description, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.According to this hair Embodiment in bright, every other implementation obtained by those of ordinary skill in the art without making creative efforts Example, shall fall within the protection scope of the present invention.

Risk model method for building up provided in an embodiment of the present invention, can be applicable in the application environment such as Fig. 1, wherein visitor Family end is communicated by network with server.Wherein, client can be, but not limited to various personal computers, notebook electricity Brain, smart phone, tablet computer and portable wearable device.Server can use the either multiple services of independent server The server cluster of device composition is realized.

In one embodiment, it as shown in Fig. 2, providing a kind of risk model method for building up, applies in Fig. 1 in this way It is illustrated, includes the following steps: for server

S10: at least one Claims Resolution information table, the corresponding Customer ID of each Claims Resolution information table and an insurance kind type are obtained.

Wherein, Claims Resolution refers to the insurer when insurance risk occurs for the insurance subject accepted insurance, and insurance contract party proposes It after the request of claim, is provided according to insurance contract, audits the juristic act of insurance indemnity.When Claims Resolution information table refers to application Claims Resolution The registration form of reporting a case to the security authorities submitted.Insurance company provides according to insurance contract, audits to Claims Resolution information table, and determine whether to Give insurance indemnity.

Specifically, be stored in the database built in server or the database being connected with server information table of insuring, Insurance kind information table and all Claims Resolution information tables that Claims Resolution occurred, information table of insuring, insurance kind information table and Claims Resolution information table storage In the different file of database.When information table of insuring refers to that insurer proposes to insure request to the insurer, the information table filled in, It include Customer ID, age, marital status, occupation, the income of insurer and source, sleep quality, education in information table of insuring Whether degree, working time section, job site, trip mode and be three high fields, insurer is according in information table of insuring Field fills in corresponding field value, these field values are the information of insurer.For example, marital status can be married or unmarried, Working time section can be morning on working day 8:00-6:00 etc..Insurance kind information table refers to the condition clause and insurance responsibility of insurance kind.Its In, the corresponding Claims Resolution accident of each Claims Resolution information table, server obtains at least one Claims Resolution information table from database, each A corresponding Customer ID and an insurance kind type in Claims Resolution information table, wherein Customer ID is the mark of insurer for identification, specifically may be used To be ID card No., unique client is able to confirm that by Customer ID.Insurance kind type is property according to Insurance Management, mesh , object and insurance regulation require and the insurance kind classification of the divisions such as historical custom.It is divided with insured object, it is corresponding Insurance kind type mainly includes personal insurance and property insurance etc..

S20: counting the corresponding first Claims Resolution frequency of each Customer ID and the corresponding second Claims Resolution frequency of each insurance kind type, Target customer ID and target insurance kind type are determined according to the first Claims Resolution frequency and the second Claims Resolution frequency.

Wherein, the first Claims Resolution frequency refers to that the frequency of Claims Resolution occurred for Customer ID, and the frequency that can be settled a claim by formula first is public Formula isIt is calculated.

The second Claims Resolution frequency refers to that the frequency of Claims Resolution occurred for insurance kind type, and the frequency formula that can be settled a claim by formula second isIt is calculated.

Target customer ID refers to the Customer ID determined according to the first Claims Resolution frequency, and target insurance kind type refers to according to the second reason Pay for the insurance kind type that the frequency determines.

Specifically, server obtains Claims Resolution information table all in database, and extracts Customer ID from Claims Resolution information table With insurance kind type, claim times statistics is carried out to each Customer ID and insurance kind type, and according to the claim times of each Customer ID It determines the corresponding first Claims Resolution frequency of each Customer ID, determines that each insurance kind type is corresponding according to the claim times of insurance kind type The second Claims Resolution frequency.Wherein, the corresponding Claims Resolution frequency of each Customer ID can it is identical can be different, the insurance kind type insured is more, right The high probability of the Claims Resolution frequency answered is bigger.It can specifically be carried out by preset threshold and the first Claims Resolution frequency and the second Claims Resolution frequency Compare, to determine target customer ID and target insurance kind type.For example, the Customer ID that the first Claims Resolution frequency is greater than preset threshold is made For target customer ID, insurance kind type of the frequency greater than preset threshold is settled a claim as target insurance kind type using second, in this way, Target customer ID and target insurance kind type are got, for subsequently through the client of target customer ID and target insurance kind type-collection spy It levies dimension and insurance kind characteristic dimension and provides convenience for high risk characteristic dimension.

S30: information table of insuring is inquired according to target customer ID, obtains client characteristics dimension, and according to target insurance kind type Insurance kind information table is inquired, insurance kind characteristic dimension is obtained.

Wherein, client characteristics dimension refers to the feature dimensions corresponding with target customer ID extracted from information table of insuring Degree.Insurance kind characteristic dimension refers to the characteristic dimension corresponding with target insurance kind type extracted from insurance kind information table.

Specifically, server can inquire database, acquisition and target after determining target customer ID based on target customer ID The corresponding information table of insuring of Customer ID, wherein same Customer ID is corresponding insure information table people from China National Investment & Guaranty Corp. information can be identical Difference, for example, the insurance kind type that insurer insures is AA, it include the information table for the AA that insures that the Customer ID is filled in database, The insurance kind type that same insurer insures is BB, the information table for the BB that insures that packet Customer ID is filled in database.Wherein, two parts of throwings The customer information for protecting same insurer in information table has been likely to occur variation, such as the age of insurer, financial revenue and expenditure and marriage Situation etc., and the identification card number of insurer and date of birth etc. do not change.Server inquires database, obtains target customer ID The corresponding information table of insuring of Claims Resolution occurred, and extracts corresponding client characteristics dimension in each information table of insuring, specifically might be used Customer information in information table of insuring is matched with the customer information in corresponding Claims Resolution information table with to be, and from information of insuring The field of successful match is extracted in table as client characteristics dimension.

For example, inquiring 10 parts of information tables of insuring in information table of insuring by target customer ID, Claims Resolution occurred for acquisition Corresponding all information tables of insuring obtain and this 4 parts Claims Resolution information tables pair if the target customer ID has 4 parts of Claims Resolution information tables The information table of insuring answered, and customer information in each information table of insuring is carried out with the customer information in corresponding Claims Resolution information table Matching, and the client characteristics dimension of successful match is extracted, client characteristics dimension specifically can be age, marital status, occupation, throwing The income of guarantor and source, sleep quality, education degree, working time section, job site, trip mode and whether be three high Etc. fields.

For another example by target insurance kind type queries to the corresponding insurance kind information table of insurance kind, by insurance kind in insurance kind information table Information is matched with insurance kind information in corresponding Claims Resolution information table, obtains the insurance kind characteristic dimension of successful match, insurance kind information It include the corresponding all insurance responsibilities of insurance kind type and insurance responsibility corresponding waiting period, segmentation condition and Claims Resolution item in table Part etc., the corresponding content of fields such as waiting period, segmentation condition, Claims Resolution condition may be the same or different in each insurance responsibility.Example Such as, insurance responsibility field is medical treatment payment in insurance kind information table, wherein waiting period field is declaration form in the phase of paying dues in medical treatment payment After coming into force three months；Claims Resolution condition field is because accident or disease are hospitalized；Amount for which loss settled field is subitem payment amount of money insurance money 1.6 ten thousand yuan, by matching insurance kind information in insurance kind information table with insurance kind information in Claims Resolution information table, obtain successful match Field as insurance kind characteristic dimension.

S40: according to client characteristics dimension and insurance kind characteristic dimension, being modeled using decision Tree algorithms, obtains risk mould Type.

Wherein, decision tree is also known as decision tree, is a prediction model, wherein what decision tree represented is object properties with A kind of mapping relations between object value, each node indicates some object (i.e. characteristic dimension) in tree, and each diverging paths The possible attribute value of some then represented, and each leaf node represent be object value (i.e. the characteristic value of characteristic dimension).

Specifically, the input of decision tree modeling is the table being made of client characteristics dimension and insurance kind characteristic dimension, In, corresponding weight is configured with to client characteristics dimension and insurance kind characteristic dimension, modeling the result is that a binary tree or multi-fork Tree.The node of binary tree is typically expressed as a logic judgment, if form is b=b_iLogic judgment, wherein b be feature because Son, b_i(attribute value) is all values of this feature factor, and the side of tree is the branch outcome of logic judgment.Decision Tree algorithms can be with Within the relatively short time to large data source make it is feasible and work well as a result, it is possible to increase risk model it is accurate Rate, and decision tree only needs once to construct, and Reusability improves the efficiency of risk model.

In step S10-S40, by obtaining information table of settling a claim, to count Customer ID and insurance kind in each Claims Resolution information table The Claims Resolution frequency of type, and target customer ID and target insurance kind type are determined according to the Claims Resolution frequency of Customer ID and insurance kind type, It is not required to largely calculate and gets target customer ID and target insurance kind type, acquisition methods are simple and quick；According to target customer ID and Target insurance kind type determines client characteristics dimension and insurance kind characteristic dimension, so as to the client characteristics dimension and insurance kind feature dimensions of extraction Degree is that the Claims Resolution higher characteristic dimension of the frequency occurs, and helps to improve the accuracy rate of modeling；It can be in phase by decision Tree algorithms Client characteristics dimension and insurance kind characteristic dimension are modeled in the short time, risk model, and decision can be quickly obtained Tree only needs primary building, and Reusability improves the service efficiency of risk model.

In one embodiment, as shown in figure 3, in step S20, that is, count each Customer ID it is corresponding first Claims Resolution the frequency and The corresponding second Claims Resolution frequency of each insurance kind type, according to the first Claims Resolution frequency and the second Claims Resolution frequency determine target customer ID and Target insurance kind type, specifically comprises the following steps:

S21: the corresponding claim times of each Customer ID and the corresponding information table quantity of insuring of Customer ID are counted, pass through first Claims Resolution frequency formula is calculated, and the corresponding first Claims Resolution frequency of each Customer ID is obtained.

Wherein, the first Claims Resolution frequency formula is

Wherein, the first claim times refer to that the number of Claims Resolution occurred in the database for Customer ID.

Specifically, server obtains all Claims Resolution information tables in database, extracts Customer ID in all Claims Resolution information tables, right Each Customer ID is counted, and the corresponding claim times of each Customer ID are obtained.Wherein, the corresponding claim times of each Customer ID The claim times that can be the corresponding each insurance kind type of same Customer ID are also possible to all insurance kinds that same Customer ID is insured The claim times of type.

Further, the corresponding claim times of each Customer ID can be the corresponding each insurance kind type of same Customer ID Claim times, for example, the insurance kind type in the Claims Resolution information table of a certain Customer ID includes that vehicle insurance and major disease insure both, The claim times that vehicle insurance and major disease insurance can be counted respectively, by by Customer ID in Customer ID and all information tables of insuring It is matched, obtains the information table of insuring with Customer ID successful match, and according to the insurance kind type and Claims Resolution in information table of insuring Insurance kind type in information table obtains the Customer ID in vehicle insurance and the major disease claim times insured and information table of insuring respectively Quantity, then by first Claims Resolution frequency formula beIt is calculated, Obtain the Customer ID vehicle insurance and the corresponding first Claims Resolution frequency of major disease insurance.

In another example server obtains 4 parts of Claims Resolutions information table corresponding with the Customer ID, it include vehicle in 4 parts of Claims Resolution information tables Danger and major disease insure both insurance kind types, are 2 by obtaining information table quantity of insuring corresponding with the Customer ID, will manage The insurance kind type paid in information table is matched with the insurance kind type in information table of insuring, when in the information table of insuring of the Customer ID Insurance kind type be vehicle insurance and the insurance kind type in 3 parts of Claims Resolution information tables is the equal successful match of vehicle insurance, obtain the Claims Resolution letter of vehicle insurance Ceasing table is 3 parts, and the corresponding claim times of Claims Resolution information table, i.e., it is 3 that insurance kind type, which is the claim times of vehicle insurance, and insurance kind type is Information table quantity of insuring is 1, then the Claims Resolution frequency that insurance kind type is vehicle insurance is calculated by formula is 3.When the Customer ID The insurance kind type in information table of insuring is great sickness insurance and the insurance kind type in 1 part of Claims Resolution information table is great sickness insurance Successful match, then it is 1 that insurance kind type, which is the claim times of great sickness insurance, and insurance kind type is that great sickness insurance is insured information Table quantity is 1, then the Claims Resolution frequency that insurance kind type is great sickness insurance is calculated by formula is 1.

Further, the corresponding claim times of each Customer ID are the Claims Resolutions for all insurance kind types that same Customer ID is insured Number.For example, getting two parts of Claims Resolution information tables by a certain Customer ID, wherein insurance kind type includes vehicle in Claims Resolution information table Danger and major disease insurance, statistics insurance kind type are total compensation number that vehicle insurance and major disease are insured, by by Customer ID and institute Customer ID in information table of insuring is matched, and obtains the information table of insuring with Customer ID successful match, and according to information of insuring Insurance kind type in insurance kind type in table and Claims Resolution information table obtains the corresponding vehicle insurance of the Customer ID and major disease insurance Claim times vehicle insurance corresponding with the Customer ID and major disease insure corresponding information table quantity of insuring, and pass through the first Claims Resolution Frequency formula isIt is calculated, obtains each client corresponding The one Claims Resolution frequency.For example, obtaining 4 parts of Claims Resolutions information table corresponding with the Customer ID by Customer ID, a Claims Resolution information table is corresponding One claim times, obtaining information table quantity of insuring corresponding with the Customer ID by the Customer ID is 2, passes through formulaThe first Claims Resolution frequency that the Customer ID is calculated is 2.Since Claims Resolution information table can insure for same Main danger or accessory risk etc. are settled a claim in information, more parts of Claims Resolution information tables occur and correspond to portion insuring information table, then the reason obtained It is higher to pay for the frequency.

S22: the corresponding claim times of each insurance kind type and the corresponding insurance kind information table quantity of insurance kind type are counted, are passed through Second Claims Resolution frequency formula is calculated, and the corresponding second Claims Resolution frequency of each insurance kind type is obtained.

Wherein, the second Claims Resolution frequency formula is

Wherein, the second claim times refer to that the number settled a claim occurred for insurance kind type in the database.

Specifically, server obtains all Claims Resolution information tables, extracts to insurance kind type in each Claims Resolution information table, unites Count the quantity of the Claims Resolution information table of each insurance kind type, the i.e. corresponding claim times of insurance kind type.Wherein, it is stored in database The corresponding insurance kind information table of each insurance kind type, the corresponding insurance kind information table of same insurance kind type pass through the second Claims Resolution frequency Formula isIt is calculated to get the Claims Resolution of each insurance kind is arrived Number is the corresponding Claims Resolution frequency of each insurance kind type.For example, server inquires database, it include 100 parts of Claims Resolutions in database Information table obtains the insurance kind type in all Claims Resolution information tables, the claim times of each insurance kind type is counted, if a certain insurance kind class The quantity of the Claims Resolution information table of type is 10, i.e., the claim times of the insurance kind type are 10, and the corresponding insurance kind of each insurance kind type Information table, then passing through formulaThe Claims Resolution frequency for obtaining the insurance kind type is 10.

S23: if the first Claims Resolution frequency is greater than the first preset threshold, the corresponding Customer ID of the first Claims Resolution frequency is determined as Target customer ID.

Wherein, the first preset threshold is set previously according to actual conditions, and preset threshold can be an integer.

Specifically, server obtains the corresponding first Claims Resolution frequency of each Customer ID, by each Customer ID corresponding first The Claims Resolution frequency is compared with the first preset threshold, when the first Claims Resolution frequency is greater than the first preset threshold, then will be greater than first The Customer ID of preset threshold is determined as target customer ID.

It further, is corresponding each according to same Customer ID when the corresponding first Claims Resolution frequency of the Customer ID of acquisition One or more Claims Resolution frequencys that the claim times of insurance kind type are calculated, then can by the claim times of each insurance kind type with First preset threshold compares, to determine target customer ID.Wherein, since the difference of insurance kind type, preset first is default Threshold value is also different, and corresponding first preset threshold of high risk insurance kind is greater than corresponding first predicted threshold value of low-risk insurance kind. For example, vehicle insurance is the insurance kind of high risk, settles a claim 5 times per capita, the first preset threshold of vehicle insurance can be set as 5；Major disease is protected Danger is low-risk insurance kind, settles a claim 2 times per capita, then the first preset threshold of major disease can be set as 2.Customer ID is corresponding The first Claims Resolution frequency the first preset threshold corresponding with insurance kind type of each insurance kind type compares, when in same Customer ID When the first Claims Resolution frequency of any insurance kind type is greater than the insurance kind type corresponding first preset threshold, then the Customer ID is determined For target customer ID.For example, a certain Customer ID has purchased vehicle insurance and major disease insurance, vehicle insurance and major disease are calculated separately out Insuring the corresponding first Claims Resolution frequency is 5 and 2, pre- by the first of the first Claims Resolution frequency of the corresponding vehicle insurance of the Customer ID and vehicle insurance If threshold value is compared, by the first Claims Resolution frequency and the first of major disease insurance of the corresponding major disease insurance of the Customer ID Preset threshold is compared, when the first Claims Resolution frequency of vehicle insurance and the first Claims Resolution frequency of major disease insurance are all larger than first in advance If when threshold value, then extracting to the Customer ID, it is determined as target customer ID.

Further, when the corresponding first Claims Resolution frequency of the Customer ID of acquisition, be insured according to same client it is all The claim times of insurance kind type calculate the first Claims Resolution frequency obtained, by the first Claims Resolution frequency with insurance kind type is subjected to group in advance Merge the first preset threshold corresponding to insurance kind type combination to compare, wherein insurance kind type combination corresponding first is preset Threshold value, which refers to, is combined insurance kind, and to the first preset threshold that each combination is configured, for example, insurance kind type be A, B, C and D, by insurance kind type be combined with each other at ABCD, ABC, ABD ..., BC and CD, each combination is pre-configured with corresponding First preset threshold.Insurance kind type in the corresponding Claims Resolution information table of same Customer ID is matched with insurance kind type combination, is obtained The first Claims Resolution frequency is greater than the Customer ID of the first preset threshold by the first preset threshold for taking the insurance kind type combination of successful match It is determined as target customer ID.For example, include that vehicle insurance A and major disease insure B in the corresponding Claims Resolution information table of a certain Customer ID, it should The first Claims Resolution frequency of Customer ID is 4, then the first Claims Resolution frequency 4 and insurance kind type combination is insured B for vehicle insurance A and major disease Corresponding first preset threshold compares, and obtains the first Claims Resolution frequency and is greater than the Customer ID of first preset threshold, and determines For target customer ID.

S24: if the second Claims Resolution frequency is greater than the second preset threshold, the corresponding insurance kind type of the second Claims Resolution frequency is determined For target insurance kind type.

Wherein, the second preset threshold is set previously according to actual conditions, and preset threshold is integer.

Specifically, server is after obtaining the corresponding second Claims Resolution frequency of each insurance kind type, by each insurance kind type pair The second Claims Resolution frequency answered is compared with the second preset threshold, when the second Claims Resolution frequency is greater than the second preset threshold, then will Insurance kind type greater than the second preset threshold is determined as target insurance kind type.Wherein, the second preset threshold is to every in database One insurance kind type is subjected to the preset threshold of claim times range setting according to insurance company.For example, insurance kind type is vehicle insurance, The information table of insuring for obtaining all vehicle insurances in database is 1000, and the maximum acceptable claim times of insurance company are information table of insuring Second preset threshold of vehicle insurance is then set as 400, obtains the claim times of vehicle insurance by the 40% of quantity, when claim times are greater than 400 When, using vehicle insurance as target insurance kind type, wherein insurance company for each maximum acceptable claim times of insurance kind type not Together, then corresponding second preset threshold of each insurance kind type is different.

In step S21-S24, the first reason of Customer ID is determined according to the claim times of Customer ID and information table quantity of insuring The frequency is paid for, determines the second claim times of insurance kind type according to the claim times of insurance kind type and insurance kind information table quantity, first The acquisition process of the frequency of settling a claim and the second Claims Resolution frequency is simple and convenient, moreover, by client characteristics dimension and insurance kind feature dimensions Degree embodies the degree that Claims Resolution risk occurs.When the first Claims Resolution frequency is greater than the first preset threshold, obtain and the first Claims Resolution frequency Corresponding Customer ID, and it is determined as target customer ID, so that the corresponding client of target customer ID got is that Claims Resolution time occurs The more crowd of number, i.e. high risk client can make so that the later use high risk client obtains corresponding client characteristics dimension The client characteristics dimension of acquisition can preferably embody the feature of high risk client, further help in the risk mould for improving and establishing The accuracy of type.When the second Claims Resolution frequency is greater than the second preset threshold, insurance kind type corresponding with the second Claims Resolution frequency is obtained, And it is determined as target insurance kind type, i.e. high risk insurance kind, for subsequently through the corresponding insurance kind feature dimensions of target insurance kind type-collection Degree can enable the insurance kind characteristic dimension obtained preferably embody the feature of high risk insurance kind, further help in raising and establish Risk model accuracy.

In one embodiment, Claims Resolution information table further includes Claims Resolution record.It not only include client in i.e. each Claims Resolution information table ID and insurance kind type, settle a claim information table in further include Claims Resolution record.Wherein, traffic injury time, accident are sent out in Claims Resolution record Radix Rehmanniae point and process, accident occurrence cause and in detail by etc. recorded, wherein to detailed by carrying out in Claims Resolution record When record, illustrate which condition in insurance kind type met because of which information generation Claims Resolution accident of client and accident of settling a claim (can be insurance responsibility corresponding waiting period, segmentation condition and Claims Resolution condition).Traffic injury time can be generation on daytime Or night, place where the accident occurred point can be insurance place, family neutralizes other places etc..

As shown in figure 4, step S30, i.e., inquire information table of insuring according to target customer ID, client characteristics dimension is obtained, and According to target insurance kind type queries insurance kind information table, insurance kind characteristic dimension is obtained, is specifically comprised the following steps:

S31: inquiring information table of insuring according to target customer ID, obtain target customer's information corresponding with target customer ID, Claims Resolution record in the corresponding Claims Resolution information table of target customer ID is subjected to fuzzy matching with target customer's information, from target customer Client characteristics dimension is extracted in information.

Wherein, target customer's information refers to customer information corresponding with target customer ID, for example, target customer ID is 18 The ID card No. 4403xxxxxxxxxxxxxx of position, finds letter of insuring by ID card No. 4403xxxxxxxxxxxxxx Table is ceased, using the customer information in information table of insuring as target customer's information.Client characteristics dimension refer to from information table of insuring The Claims Resolution record successful field of fuzzy matching in the customer information and Claims Resolution information table of middle acquisition.

Specifically, the Claims Resolution record in the target customer's information and Claims Resolution information table in information table of insuring is carried out fuzzy Match, gets effective client characteristics dimension.In fuzzy matching, can be used real using like and % in SQL search statement Existing fuzzy matching carries out fuzzy matching by lookup function etc..Firstly, server can be calculated by going stop words to extract Method etc. handles the Claims Resolution record in Claims Resolution information table, particular by the nothing gone in deactivated word algorithm rejecting Claims Resolution record The word or symbol of physical meaning, such as, and symbol, will in Claims Resolution record remaining character as keyword.Then, Like condition is set based on the keyword in Claims Resolution record, the character of any type and random length is matched by SQL statement, it is right In Chinese Fields, two percentage signs (% ... %) can be used to identify, the information table for example, select keyword from insures Where keyword like% condition %, from the field obtained in information table of insuring with keyword matches in Claims Resolution record.

For example, referring to that time of origin, scene and detailed process are certain year in such a month, and on such a day morning nine in Claims Resolution record Point knocks into the back because of drunk driving in the main road XX ... etc., by go to deactivate word algorithm obtain Claims Resolution record in all keyword, example Such as, client believes in " 9 points of morning ", " drunk driving ", " main road XX " and " knocking into the back ", the keyword that will acquire and information table of insuring Breath carry out fuzzy matching, wherein in information table of insuring target customer's information include the working time, job site, home address and Fields and the corresponding field values such as smoking and drinking habit, can be matched to the working time when carrying out fuzzy matching by " 9 points of morning " Section field can be matched to smoking and drinking by " drunk driving " and be accustomed to field, can be matched to job site or family by " main road XX " Front yard address field is not matched to corresponding field by " knocking into the back ", when being matched to corresponding word in target customer's information Section then determines keyword and the success of target customer's information matches in Claims Resolution information table, by the successful match from information table of insuring Field extracts, and is determined as client characteristics dimension.

S32: according to target insurance kind type queries insurance kind information table, target insurance kind letter corresponding with target insurance kind type is obtained Claims Resolution record in the corresponding Claims Resolution information table of target insurance kind type is carried out fuzzy matching with target insurance kind information, from mesh by breath It marks and extracts insurance kind characteristic dimension in insurance kind information.

Wherein, target insurance kind information refers to insurance kind information corresponding with target insurance kind type.For example, insurance kind type is vehicle Danger, finds the insurance kind information table of vehicle insurance, wherein number corresponding with target insurance kind type in insurance kind information table in the database According to for target insurance kind information.Insurance kind characteristic dimension, which refers to, to be recorded from what is extracted in insurance kind information table with the Claims Resolution in Claims Resolution information table The successful field of fuzzy matching.

Specifically, it when server gets target insurance kind type, searches in the database corresponding with target insurance kind type Insurance kind information table, and obtain the target insurance kind information in insurance kind information table, by insurance kind information table target insurance kind information with Claims Resolution record in Claims Resolution information table carries out fuzzy matching, gets effective insurance kind characteristic dimension.Wherein, in fuzzy matching, It can be used and realize fuzzy matching using like and % in SQL search statement or carry out fuzzy by lookup function etc. Match.Firstly, server can be by going stop words extraction algorithm etc. to handle Claims Resolution record in Claims Resolution information table, specifically Reject word or symbol without physical meaning in Claims Resolution record by removing to deactivate word algorithm, such as, and symbol, will settle a claim Remaining character is as keyword in record.Then, based on the keyword in Claims Resolution record, like condition is set, when passing through SQL The character of statement matching any type and random length two percentage signs (% ... %) can be used to identify Chinese Fields, For example, select keyword from insurance kind information table where keyword like% condition %, obtains and is managed from insurance kind information table The field that the keyword in record matches is paid for, as client characteristics dimension.

For example, referring to " traffic accident occurring, lead to other side driver's disability " in Claims Resolution record, word algorithm is deactivated by going Keyword " traffic accident " and " leading to other side driver's disability " in Claims Resolution record are obtained, it will be in keyword and insurance kind information table Target insurance kind information carries out fuzzy matching.It wherein, include specific accident/injury insurance, specific unexpected injury in target insurance kind information The Claims Resolution condition of insurance is that insurance motor vehicle occurs road traffic accident victim is caused (not include this vehicle personnel and insured People) personal injury, property loss, the mandatory liability insurance to give compensation in liability limit.Pass through keyword " traffic thing Therefore " and " leading to other side driver's disability " and the Claims Resolution condition of specific accident/injury insurance match, then extract specific accidental wound Specific accident/injury insurance corresponding insurance motor vehicle road traffic accident to cause victim (not including this by evil insurance occurs Vehicle personnel and insurant) personal injury, property loss, the mandatory liability insurance conduct to give compensation in liability limit Insurance kind characteristic dimension.

In step S31-S32, by the way that the Claims Resolution record in target customer's information and Claims Resolution information table is carried out fuzzy matching, Client characteristics dimension is obtained from information table of insuring；Target insurance kind information and the Claims Resolution record in Claims Resolution information table are obscured Matching, from insurance kind information table obtain insurance kind characteristic dimension, so as to get high risk client client characteristics dimension and high wind The insurance kind characteristic dimension of insurance kind can preferably embody high risk by the client characteristics dimension and insurance kind characteristic dimension of acquisition The feature of client and high risk insurance kind further help in the accuracy for improving the risk model established.

In one embodiment, as shown in figure 5, step S40 is used that is, according to client characteristics dimension and insurance kind characteristic dimension Decision Tree algorithms are modeled, and are obtained risk model, are specifically comprised the following steps:

Step S41: corresponding weight is respectively configured to client characteristics dimension and insurance kind characteristic dimension.

Specifically, to the corresponding characteristic value of each client characteristics dimension (in information table of insuring with client characteristics dimension pair The field value answered) the corresponding weight of configuration.For example, client characteristics dimension is smoking and drinking habit, smoking and drinking is accustomed to this Characteristic value (such as smoking and drinking, non-smoking drink, smoke not drinking and drink with non-smoking) in client characteristics dimension is respectively configured Corresponding weight.Wherein, to the corresponding weight of eigenvalue assignment mainly according to all client characteristics in same information table of insuring Importance in the dimension or importance of different characteristic value carries out configuring corresponding weight in same client characteristics dimension.Example Such as, insurance kind type is major disease, and the client characteristics dimension of extraction includes job site, working time, smoking and drinking, diet work Breath and the characteristic dimensions such as physical training, if mainly causing major disease by smoking and drinking and working environment, by smoking and drinking and Working environment dimension configures higher weight, and higher weight is arranged in other characteristic dimensions.For another example it is special that three high clients occur It is more than the feature dimensions such as 30 years old youth, eating habit, working environment, working time and job site that sign dimension, which is distributed in the age, This characteristic dimension of eating habit is then configured higher weight since three is high mainly as caused by eating habit by degree.Wherein, The weight of each client characteristics dimension is the sum of corresponding characteristic value weight, for example, marital status characteristic dimension, to characteristic value Wedding and the corresponding weight of unmarried configuration, weight=married weight+unmarried weight of marital status characteristic dimension.

Further, to the corresponding characteristic value of each insurance kind characteristic dimension (i.e. in insurance kind information table with insurance kind characteristic dimension Corresponding field value) the corresponding weight of configuration.When to the corresponding weight of eigenvalue assignment in insurance kind characteristic dimension, according to same Claim times configure corresponding weight in insurance kind type.For example, field is personal accidental death and injury insurance, it is common unexpected by feature value division Injury danger and specific personal accidental death and injury insurance, if the number that the common personal accidental death and injury insurance occurred is settled a claim is more, common accidental wound The weight that evil nearly configures configures higher weight relative to specific personal accidental death and injury insurance.

Step S42: according to client characteristics dimension and the corresponding weight of insurance kind characteristic dimension, feature weight table is obtained.

Wherein, feature weight table refers to the table that corresponding weight is configured according to client characteristics dimension and insurance kind characteristic dimension.

Specifically, each client characteristics dimension and corresponding characteristic value are counted, and count in each insurance kind characteristic dimension and Corresponding characteristic value, using client characteristics dimension and insurance kind characteristic dimension as the characteristic dimension in feature weight table, and by client The corresponding characteristic value of characteristic dimension and the corresponding characteristic value of insurance kind characteristic dimension are as the characteristic value in feature weight table.For example, Client characteristics dimension is " smoking and drinking habit ", the corresponding characteristic value of client characteristics dimension " smoking and drinking ", then weighing in feature " smoking and drinking habit " is used as characteristic dimension in weight table, the corresponding characteristic value of client characteristics dimension " smoking and drinking " is used as feature Value.According to the weight configured in advance to client characteristics dimension and insurance kind characteristic dimension, feature weight table is generated, in characteristic dimension table In, show the weight of each client characteristics dimension and the corresponding each characteristic value of insurance kind characteristic dimension.For example, in feature weight table In, in this characteristic dimension of smoking and drinking, display smoking and drinking, that non-smoking is drunk, smoking does not drink and smoking does not drink etc. is special Value indicative, moreover, smoking and drinking, non-smoking are drunk, smoking not drinking respectively corresponds different power with the non-smoking characteristic values such as drink Weight.

Step S43: modeling feature weight table using ID3 algorithm, obtains risk model.

Wherein, ID3 (Iterative Dichotomiser 3, iteration binary tree) algorithm is that one kind is used to construct decision The algorithm of tree, it carries out Attributions selection according to information gain.In every single-step iteration of algorithm, traverse in feature weight table Each of not used characteristic dimension (characteristic dimension refers to client characteristics dimension or insurance kind characteristic dimension), calculate this feature dimension Entropy or information gain, therefrom select root node of the characteristic dimension with minimum entropy or maximum information gain as binary tree, The characteristic value in characteristic dimension is divided into different attribute values with selected characteristic dimension again, and (such as age characteristics dimension is less than 30 In year, the age, the age was greater than 50 years old between 30~50 years old), continue to carry out Recursion process to feature weight table by ID3 algorithm, The attribute that do not select before only considering every time is completed until binary tree is established.Specifically, it using ID3 algorithm, is weighed according to feature Characteristic dimension establishes binary tree or multiway tree in weight table, according to the information gain for calculating each characteristic dimension, determines each feature The level of dimension and corresponding weight in binary tree or multiway tree, in the present embodiment, due to the feature dimensions in feature weight table Degree is the validity feature dimension extracted from insure information table and insurance kind information table, is not required to consider the problems of beta pruning, wherein effectively Characteristic dimension refers to by successful with Keywords matching in Claims Resolution information table, thus the client characteristics dimension and insurance kind feature that obtain Dimension.

In step S41-S43, by configuring corresponding weight to client characteristics dimension and insurance kind characteristic dimension, feature is established Weight table models feature weight table by ID3 algorithm, to obtain risk model, for subsequently through weight calculation whether Convenience is provided for high risk information, feature weight table is modeled by ID3 algorithm, can be got in a relatively short period of time One binary tree, and binary tree only needs once to construct, and Reusability improves the service efficiency of risk model.

In one embodiment, it as shown in fig. 6, step S43, i.e., model feature weight table using ID3 algorithm, obtains Risk model specifically comprises the following steps:

S431: traversal feature weight table passes through the phase according to client characteristics dimension in feature weight table and insurance kind characteristic dimension Prestige value formula, comentropy formula and information gain formula are calculated, and the information gain of each client characteristics dimension and every is obtained The information gain of one insurance kind characteristic dimension.

Wherein, desired value formula isComentropy formula is Information gain formula is Gain (x, t)=Info (t)-Info (x, t), p_iIt is i-th of client characteristics dimension or insurance kind feature dimensions The probability of degree, Info (t) are the desired values of client characteristics dimension or insurance kind characteristic dimension, and m is that client characteristics dimension or insurance kind are special The quantity of dimension is levied, t is client characteristics dimension or insurance kind characteristic dimension, and x is all client characteristics dimensions or insurance kind characteristic dimension Corresponding weighted mean, x_iIt is i-th of client characteristics dimension or the corresponding weight of insurance kind characteristic dimension, Info (x, t) is visitor The comentropy of family characteristic dimension or insurance kind characteristic dimension, Gain (x, t) are the information of client characteristics dimension or insurance kind characteristic dimension Gain.

Specifically, it is by desired value formulaIt is special to client each in characteristic dimension table x Sign dimension t or insurance kind characteristic dimension t carries out expectation value information, obtains the expectation of the client characteristics dimension t or insurance kind characteristic dimension t Value Info (t)；It is by comentropy formula againTo client characteristics dimension t or insurance kind characteristic dimension The desired value Info (t) of t is calculated, and the comentropy Info of each client characteristics dimension t or insurance kind characteristic dimension t are calculated (x,t)；It then, is that Gain (x, t)=Info (t)-Info (x, t) calculates each client characteristics dimension by information gain formula The information gain Gain (x, t) for spending t or insurance kind characteristic dimension t, determines node according to information gain Gain (x, t) so as to subsequent.

S432: the root section for having the client characteristics dimension or insurance kind characteristic dimension of maximum information gain as decision tree is chosen Point, and using client characteristics dimension or the corresponding weight of insurance kind characteristic dimension as the weight of root node.

Specifically, by the information gain Gain (x, t) and each insurance kind spy of each client characteristics dimension t being calculated The information gain Gain (x, t) of dimension t is levied, the maximum information gain in client characteristics dimension t or insurance kind characteristic dimension t is obtained Gain (x, t), using maximum information gain G ain (x, t) corresponding client characteristics dimension t or insurance kind characteristic dimension t as decision tree Root node.For example, including that smoking and drinking is accustomed to characteristic dimension, age characteristics dimension, ergonomic features in feature weight table Dimension, the information gain for being accustomed to characteristic dimension by the way that smoking and drinking is calculated are greater than the information gain of age characteristics dimension, inhale Cigarette, which is drunk, is accustomed to information gain of the information gain greater than ergonomic features dimension of characteristic dimension, then is accustomed to smoking and drinking special Root node of the dimension as decision tree is levied, using the corresponding weighted value of smoking and drinking habit characteristic dimension in feature weight table as root The weight of node, and using the weight of the corresponding each characteristic value of smoking and drinking habit characteristic dimension in feature weight table as root section The weight of each diverging paths of point.For example, root node is habit of smoking and drink, it is divided into smoking and drinking, non-smoking is drunk, is smoked not It drinks and smoking four diverging paths that do not drink according to weight preconfigured in feature weight table determines each diverging paths Weight.

S433: the root node based on decision tree repeats the information gain for obtaining each client characteristics dimension and each The step of information gain of insurance kind characteristic dimension, successively determines the decision tree second layer to each node of the last layer, until each branch The corresponding characteristic value of lower lowest level node is all same client characteristics dimension or insurance kind characteristic dimension, obtains risk model.

Specifically, it is determined that, according to the characteristic value of the corresponding characteristic dimension of each branch node, being generated respectively every after root node The decision tree branches of a branch node, when determining lower layer of decision tree branches of branch node, and will be objective in feature weight table Family characteristic dimension t or insurance kind characteristic dimension t subtracts fixed client characteristics dimension t or insurance kind characteristic dimension t, and will subtract Candidate feature dimension of the determining client characteristics dimension t or insurance kind characteristic dimension t as the branch successively determines decision tree second Layer is to the corresponding characteristic dimension of each node of the last layer, when the corresponding characteristic value of lowest level node is all same feature under each branch Dimension obtains risk model.For example, lowest level node is age characteristics dimension, corresponding characteristic value was the age less than 30 years old, year Age, the age was greater than 50 years old for same characteristic dimension between 30~50 years old, then obtained risk model.

In step S431-S433, characteristic dimension table is modeled using ID3 algorithm, that is, establishes the growth of decision tree Journey.In decision tree in growth, by calculating the information gain of each characteristic dimension, the corresponding feature of maximum information gain is chosen The root node that dimension and corresponding characteristic value are grown as optimal characteristics and optimal cut-off as decision tree, then proceedes to iteration Until the corresponding characteristic value of lowest level node is all same characteristic dimension, risk model is obtained, by ID3 algorithm relatively short And the decision tree that works well feasible to data creation in feature weight table in time, and decision tree only needs once to construct, Reusability improves the service efficiency of risk model.

Risk matching process provided in an embodiment of the present invention, can be applicable in the application environment such as Fig. 1, wherein client It is communicated by network with server.Wherein, client can be, but not limited to various personal computers, laptop, intelligence It can mobile phone, tablet computer and portable wearable device.Server can use independent server either multiple server groups At server cluster realize.

In one embodiment, as shown in fig. 7, providing a kind of risk matching process, which further includes Following steps:

S71: information to be matched of insuring is obtained.

Wherein, when information to be matched of insuring refers to that client buys insurance, the information of insuring filled in.The information to be matched of insuring Including but not limited to Customer ID, insurance kind type, working environment, smoking and drinking habit, age, marital status, occupation, insurer Income and source, sleep quality, education degree and working time section etc..

S72: feature extraction is carried out to information to be matched of insuring, obtains target customer's characteristic value and target insurance kind feature respectively Value.

Specifically, server extracts client characteristics dimension and insurance kind characteristic dimension from information to be matched of insuring, and will extract The corresponding customer information of client characteristics dimension and the corresponding insurance kind information of insurance kind characteristic dimension be determined as target customer's characteristic value With target insurance kind characteristic value.Wherein, target customer's characteristic value refers to the characteristic value from information extraction to be matched of insuring, i.e., to be matched The corresponding field value of customer information fields in information of insuring.Target insurance kind feature refers to the danger extracted from information to be matched of insuring The corresponding characteristic value of seed type, i.e., the corresponding field value of insurance responsibility in insurance kind type, field value can be by waiting period, segmentation Condition and Claims Resolution condition etc. are constituted.

S73: the risk model obtained using risk model method for building up, to target customer's characteristic value and target insurance kind feature Value carries out risk matching, obtains risk assessment value.

Wherein, risk model is the model obtained using above-mentioned risk model method for building up, using the risk model by mesh Mark client characteristics value and target insurance kind characteristic value are matched with the characteristic value of each node of risk model, are obtained in risk model The node of successful match, according to the wind of the corresponding weight calculation of the characteristic value information to be matched of insuring of node each in risk model Dangerous assessed value, so that the risk assessment value accuracy rate obtained is high.Specifically, risk assessment value calculation formula can be passed throughI=0,1,2 ... m calculates the risk assessment value of information to be matched of insuring, wherein w_iRefer to The weight of the node i of successful match.

S74: if risk assessment value is greater than default value-at-risk, information to be matched of insuring is high risk information.

Wherein, default value-at-risk is to carry out according to the actual situation preset, can be a percentage.

Specifically, the risk assessment value for the information to be matched of insuring being calculated is compared with default value-at-risk, when When risk assessment value is greater than default value-at-risk, then the information to be matched of insuring is high risk information；Be less than when risk assessment value or When equal to default value-at-risk, then the information to be matched of insuring is low-risk information.

Step S71-S74, by obtaining information to be matched of insuring, with will pass through risk model to it is to be matched insure information into Row risk assessment, energy quick obtaining risk assessment value, what the artificial verification of reduction generated verifies what insufficient or subjective judgement was made mistakes, Determine that band matches whether information of insuring is high risk information according to risk assessment value, to assist insurance company's quick lock in high risk User, to take timely measure.

It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit It is fixed.

In one embodiment, a kind of risk model is provided and establishes device, which establishes device and above-described embodiment Risk method for establishing model corresponds.As shown in figure 8, it includes data obtaining module 10, mesh that the risk model, which establishes device, Mark information determination module 20, characteristic dimension obtain module 30 and risk model establishes module 40.Each functional module is described in detail such as Under:

Data obtaining module 10, for obtaining at least one Claims Resolution information table, the corresponding Customer ID of each Claims Resolution information table With an insurance kind type.

Target information determining module 20, for counting the corresponding first Claims Resolution frequency of each Customer ID and each insurance kind type The corresponding second Claims Resolution frequency determines target customer ID and target insurance kind class according to the first Claims Resolution frequency and the second Claims Resolution frequency Type.

Characteristic dimension obtains module 30, for insuring information table according to target customer ID inquiry, obtains client characteristics dimension, And according to target insurance kind type queries insurance kind information table, insurance kind characteristic dimension is obtained.

Risk model establishes module 40, is used for according to client characteristics dimension and insurance kind characteristic dimension, using decision Tree algorithms It is modeled, obtains risk model.

Preferably, target information determining module 20 includes that the first Claims Resolution Claims Resolution frequency of frequency computing unit 21, second calculates Unit 22, target customer ID determination unit 23 and target insurance kind type determining units 24.

First Claims Resolution frequency computing unit 21, it is corresponding for counting the corresponding claim times of each Customer ID and Customer ID It insures information table quantity, is calculated by the first Claims Resolution frequency formula, obtain the corresponding first Claims Resolution frequency of each Customer ID.

Wherein, the first Claims Resolution frequency formula is

Second Claims Resolution frequency computing unit 22, for counting the corresponding claim times of each insurance kind type and insurance kind type pair The insurance kind information table quantity answered is calculated by the second Claims Resolution frequency formula, obtains corresponding second reason of each insurance kind type Pay for the frequency.

Wherein, the second Claims Resolution frequency formula is

Target customer ID determination unit 23, if being greater than the first preset threshold for the first Claims Resolution frequency, by the first Claims Resolution The corresponding Customer ID of the frequency is determined as target customer ID.

Target insurance kind type determining units 24, if being greater than the second preset threshold for the second Claims Resolution frequency, by the second reason It pays for the corresponding insurance kind type of the frequency and is determined as target insurance kind type.

Preferably, it includes that client characteristics dimension acquiring unit 31 and insurance kind characteristic dimension obtain that characteristic dimension, which obtains module 30, Unit 32.

Client characteristics dimension acquiring unit 31 obtains and target visitor for inquiring information table of insuring according to target customer ID The corresponding target customer's information of family ID, by the Claims Resolution record and target customer's information in the corresponding Claims Resolution information table of target customer ID Fuzzy matching is carried out, client characteristics dimension is extracted from target customer's information.

Insurance kind characteristic dimension acquiring unit 32, for according to target insurance kind type queries insurance kind information table, acquisition and target The corresponding target insurance kind information of insurance kind type, by target insurance kind type it is corresponding Claims Resolution information table in Claims Resolution record and target danger Kind information carries out fuzzy matching, and insurance kind characteristic dimension is extracted from target insurance kind information.

Preferably, it includes weight configuration unit 41, feature weight table acquiring unit 42 and wind that risk model, which establishes module 40, Dangerous model foundation unit 43.

Weight configuration unit 41, for corresponding weight to be respectively configured to client characteristics dimension and insurance kind characteristic dimension.

Feature weight table acquiring unit 42, for obtaining according to client characteristics dimension and the corresponding weight of insurance kind characteristic dimension Take feature weight table.

Risk model establishes unit 43, for modeling using ID3 algorithm to feature weight table, obtains risk model.

Preferably, it includes information gain computation subunit 431, root node acquisition subelement that risk model, which establishes unit 43, 432 and risk model obtain subelement 433.

Information gain computation subunit 431, for traversing feature weight table, according to client characteristics dimension in feature weight table It with insurance kind characteristic dimension, is calculated by desired value formula, comentropy formula and information gain formula, it is special to obtain each client Levy the information gain of dimension and the information gain of each insurance kind characteristic dimension.

Root node obtains subelement 432, for choosing client characteristics dimension or insurance kind feature with maximum information gain Root node of the dimension as decision tree, and using client characteristics dimension or the corresponding weight of insurance kind characteristic dimension as the power of root node Weight.

Risk model obtains subelement 433, for the root node based on decision tree, repeats and obtains each client characteristics The step of information gain of the information gain of dimension and each insurance kind characteristic dimension, successively determine the decision tree second layer to last Each node of layer, until the corresponding characteristic value of lowest level node is all same client characteristics dimension or insurance kind feature dimensions under each branch Degree obtains risk model.

The specific of device, which is established, about risk model limits the limit that may refer to above for risk model method for building up Fixed, details are not described herein.Above-mentioned risk model establish the modules in device can fully or partially through software, hardware and its Combination is to realize.Above-mentioned each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also be with It is stored in the memory in computer equipment in a software form, in order to which processor calls the above modules of execution corresponding Operation.

In one embodiment, a kind of computer equipment is provided, which can be server, internal structure Figure can be as shown in Figure 9.The computer equipment includes processor, the memory, network interface sum number connected by system bus According to library.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory of the computer equipment includes Non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating The database of machine equipment is for storing Claims Resolution information table, information of insuring and insurance kind information table etc..The network of the computer equipment connects Mouth with external terminal by network connection for being communicated.To realize a kind of risk mould when the computer program is executed by processor Type method for building up or risk matching process.

In one embodiment, a kind of computer equipment is provided, including memory, processor and storage are on a memory simultaneously The computer program that can be run on a processor, processor performs the steps of when executing computer program obtains at least one Claims Resolution information table, the corresponding Customer ID of each Claims Resolution information table and an insurance kind type；Count corresponding first reason of each Customer ID The frequency and the corresponding second Claims Resolution frequency of each insurance kind type are paid for, target is determined according to the first Claims Resolution frequency and the second Claims Resolution frequency Customer ID and target insurance kind type；Information table of insuring is inquired according to target customer ID, obtains client characteristics dimension, and according to target Insurance kind type queries insurance kind information table obtains insurance kind characteristic dimension；According to client characteristics dimension and insurance kind characteristic dimension, using certainly Plan tree algorithm is modeled, and risk model is obtained.

In one embodiment, the corresponding Claims Resolution of each Customer ID of statistics is performed the steps of when processor executes computer program Number and the corresponding information table quantity of insuring of Customer ID are calculated by the first Claims Resolution frequency formula, obtain each Customer ID pair The first Claims Resolution frequency answered；Wherein, the first Claims Resolution frequency formula is The corresponding claim times of each insurance kind type and the corresponding insurance kind information table quantity of insurance kind type are counted, the second Claims Resolution frequency is passed through Formula is calculated, and the corresponding second Claims Resolution frequency of each insurance kind type is obtained；Wherein, the second Claims Resolution frequency formula isIf the first Claims Resolution frequency is greater than the first preset threshold, will The corresponding Customer ID of the first Claims Resolution frequency is determined as target customer ID；If the second Claims Resolution frequency is greater than the second preset threshold, will The corresponding insurance kind type of the second Claims Resolution frequency is determined as target insurance kind type.

In one embodiment, it further includes Claims Resolution that Claims Resolution information table is performed the steps of when processor executes computer program Record；Information table of insuring is inquired according to target customer ID, obtains target customer's information corresponding with target customer ID, by target visitor Claims Resolution record and target customer's information in the corresponding Claims Resolution information table of family ID carry out fuzzy matching, mention from target customer's information Take client characteristics dimension；According to target insurance kind type queries insurance kind information table, target danger corresponding with target insurance kind type is obtained Claims Resolution record in the corresponding Claims Resolution information table of target insurance kind type is carried out fuzzy matching with target insurance kind information by kind information, Insurance kind characteristic dimension is extracted from target insurance kind information.

In one embodiment, it performs the steps of when processor executes computer program to client characteristics dimension and insurance kind Corresponding weight is respectively configured in characteristic dimension；According to client characteristics dimension and the corresponding weight of insurance kind characteristic dimension, feature is obtained Weight table；Feature weight table is modeled using ID3 algorithm, obtains risk model.

In one embodiment, traversal feature weight table is performed the steps of when processor executes computer program, according to spy Client characteristics dimension and insurance kind characteristic dimension in weight table are levied, desired value formula, comentropy formula and information gain formula are passed through It is calculated, obtains the information gain of each client characteristics dimension and the information gain of each insurance kind characteristic dimension；Wherein, it is expected that Value formula isComentropy formula isInformation gain formula is Gain (x, t)=Info (t)-Info (x, t), p_iIt is the probability of i-th of client characteristics dimension or insurance kind characteristic dimension, Info (t) be client characteristics dimension or insurance kind characteristic dimension desired value, m is the quantity of client characteristics dimension or insurance kind characteristic dimension, t It is client characteristics dimension or insurance kind characteristic dimension, x is all client characteristics dimensions or the corresponding weighted average of insurance kind characteristic dimension Value, x_iIt is i-th of client characteristics dimension or the corresponding weight of insurance kind characteristic dimension, Info (x, t) is client characteristics dimension or danger The comentropy of kind characteristic dimension, Gain (x, t) is the information gain of client characteristics dimension or insurance kind characteristic dimension；Choosing has most Root node of the client characteristics dimension or insurance kind characteristic dimension of big information gain as decision tree, and by client characteristics dimension or danger Weight of the corresponding weight of kind characteristic dimension as root node；Root node based on decision tree repeats and obtains each client The step of information gain of the information gain of characteristic dimension and each insurance kind characteristic dimension, successively determines the decision tree second layer to most Each node of later layer, until the corresponding characteristic value of lowest level node is all same client characteristics dimension or insurance kind feature under each branch Dimension obtains risk model.

In one embodiment, a kind of computer readable storage medium is provided, computer program, computer are stored thereon with At least one Claims Resolution information table of acquisition, each Claims Resolution information table corresponding visitor are performed the steps of when program is executed by processor Family ID and an insurance kind type；Count the corresponding first Claims Resolution frequency of each Customer ID and corresponding second Claims Resolution of each insurance kind type The frequency determines target customer ID and target insurance kind type according to the first Claims Resolution frequency and the second Claims Resolution frequency；According to target customer ID inquires information table of insuring, and obtains client characteristics dimension, and according to target insurance kind type queries insurance kind information table, it is special to obtain insurance kind Levy dimension；It according to client characteristics dimension and insurance kind characteristic dimension, is modeled using decision Tree algorithms, obtains risk model.

In one embodiment, each Customer ID of statistics corresponding reason is performed the steps of when computer program is executed by processor Number and the corresponding information table quantity of insuring of Customer ID are paid for, is calculated by the first Claims Resolution frequency formula, obtains each Customer ID The corresponding first Claims Resolution frequency；Wherein, the first Claims Resolution frequency formula is The corresponding claim times of each insurance kind type and the corresponding insurance kind information table quantity of insurance kind type are counted, the second Claims Resolution frequency is passed through Formula is calculated, and the corresponding second Claims Resolution frequency of each insurance kind type is obtained；Wherein, the second Claims Resolution frequency formula isIf the first Claims Resolution frequency is greater than the first preset threshold, will The corresponding Customer ID of the first Claims Resolution frequency is determined as target customer ID；If the second Claims Resolution frequency is greater than the second preset threshold, will The corresponding insurance kind type of the second Claims Resolution frequency is determined as target insurance kind type.

In one embodiment, it further includes reason that Claims Resolution information table is performed the steps of when computer program is executed by processor Pay for record；Information table of insuring is inquired according to target customer ID, target customer's information corresponding with target customer ID is obtained, by target Claims Resolution record and target customer's information in the corresponding Claims Resolution information table of Customer ID carry out fuzzy matching, from target customer's information Extract client characteristics dimension；According to target insurance kind type queries insurance kind information table, target corresponding with target insurance kind type is obtained Claims Resolution record in the corresponding Claims Resolution information table of target insurance kind type is carried out fuzzy with target insurance kind information by insurance kind information Match, insurance kind characteristic dimension is extracted from target insurance kind information.

In one embodiment, it is performed the steps of when computer program is executed by processor to client characteristics dimension and danger Corresponding weight is respectively configured in kind characteristic dimension；According to client characteristics dimension and the corresponding weight of insurance kind characteristic dimension, obtain special Levy weight table；Feature weight table is modeled using ID3 algorithm, obtains risk model.

In one embodiment, traversal feature weight table is performed the steps of when computer program is executed by processor, according to Client characteristics dimension and insurance kind characteristic dimension in feature weight table, it is public by desired value formula, comentropy formula and information gain Formula is calculated, and the information gain of each client characteristics dimension and the information gain of each insurance kind characteristic dimension are obtained；Wherein, the phase Prestige value formula isComentropy formula isInformation gain formula For Gain (x, t)=Info (t)-Info (x, t), p_iIt is the probability of i-th of client characteristics dimension or insurance kind characteristic dimension, Info (t) be client characteristics dimension or insurance kind characteristic dimension desired value, m is the quantity of client characteristics dimension or insurance kind characteristic dimension, t It is client characteristics dimension or insurance kind characteristic dimension, x is all client characteristics dimensions or the corresponding weighted average of insurance kind characteristic dimension Value, x_iIt is i-th of client characteristics dimension or the corresponding weight of insurance kind characteristic dimension, Info (x, t) is client characteristics dimension or danger The comentropy of kind characteristic dimension, Gain (x, t) is the information gain of client characteristics dimension or insurance kind characteristic dimension；Choosing has most Root node of the client characteristics dimension or insurance kind characteristic dimension of big information gain as decision tree, and by client characteristics dimension or danger Weight of the corresponding weight of kind characteristic dimension as root node；Root node based on decision tree repeats and obtains each client The step of information gain of the information gain of characteristic dimension and each insurance kind characteristic dimension, successively determines the decision tree second layer to most Each node of later layer, until the corresponding characteristic value of lowest level node is all same client characteristics dimension or insurance kind feature under each branch Dimension obtains risk model.

In one embodiment, a kind of risk coalignment is provided, the risk coalignment and above-described embodiment risk Method of completing the square corresponds.As shown in Figure 10, the risk coalignment include data obtaining module 71, characteristics extraction module 72, Risk matching module 73 and high risk information object information determination module 74.Detailed description are as follows for each functional module:

Data obtaining module 71, for obtaining information to be matched of insuring.

It is special to obtain target customer for carrying out feature extraction to information to be matched of insuring respectively for characteristics extraction module 72 Levy dimension and target insurance kind dimension.

Risk matching module 73, the risk model for being obtained using risk model method for building up, to target customer's feature Dimension and target insurance kind dimension carry out risk matching, obtain risk assessment value.

High risk information object information determination module 74, it is to be matched if being greater than default value-at-risk for risk assessment value Information of insuring is high risk information.

Specific about risk coalignment limits the restriction that may refer to above for risk matching process, herein not It repeats again.Modules in above-mentioned risk coalignment can be realized fully or partially through software, hardware and combinations thereof.On Stating each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also store in a software form In memory in computer equipment, the corresponding operation of the above modules is executed in order to which processor calls.

In one embodiment, a kind of computer equipment is provided, including memory, processor and storage are on a memory simultaneously The computer program that can be run on a processor, processor performs the steps of when executing computer program obtains throwing to be matched Breath information-preserving；Feature extraction is carried out to information to be matched of insuring, obtains target customer's characteristic value and target insurance kind characteristic value respectively；It adopts The risk model obtained with risk model method for building up carries out risk to target customer's characteristic value and target insurance kind characteristic value Match, obtains risk assessment value；If risk assessment value is greater than default value-at-risk, information to be matched of insuring is high risk information.

In one embodiment, a kind of computer readable storage medium is provided, computer program, computer are stored thereon with It is performed the steps of when program is executed by processor and obtains information to be matched of insuring；Feature is carried out to information to be matched of insuring to mention It takes, obtains target customer's characteristic value and target insurance kind characteristic value respectively；The risk model obtained using risk model method for building up, Risk matching is carried out to target customer's characteristic value and target insurance kind characteristic value, obtains risk assessment value；If risk assessment value is greater than Default value-at-risk, then information to be matched of insuring is high risk information.

Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with Instruct relevant hardware to complete by computer program, computer program to can be stored in a non-volatile computer readable It takes in storage medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein, this Shen Please provided by any reference used in each embodiment to memory, storage, database or other media, may each comprise Non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM (PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms, Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..

It is apparent to those skilled in the art that for convenience of description and succinctly, only with above-mentioned each function Can unit, module division progress for example, in practical application, can according to need and by above-mentioned function distribution by different Functional unit, module are completed, i.e., the internal structure of device are divided into different functional unit or module, to complete above description All or part of function.

The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although with reference to the foregoing embodiments Invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each implementation Technical solution documented by example is modified or equivalent replacement of some of the technical features；And these modification or Replacement, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution should all include Within protection scope of the present invention.

Claims

1. a kind of risk model method for building up characterized by comprising

The corresponding first Claims Resolution frequency of each Customer ID and the corresponding second Claims Resolution frequency of each insurance kind type are counted, Target customer ID and target insurance kind type are determined according to the first Claims Resolution frequency and the second Claims Resolution frequency；

Information table of insuring is inquired according to the target customer ID, obtains client characteristics dimension, and according to the target insurance kind type Insurance kind information table is inquired, insurance kind characteristic dimension is obtained；

It according to the client characteristics dimension and the insurance kind characteristic dimension, is modeled using decision Tree algorithms, obtains risk mould Type.

2. risk model method for building up as described in claim 1, which is characterized in that each Customer ID of statistics is corresponding The first Claims Resolution frequency and the corresponding second Claims Resolution frequency of each insurance kind type, according to the first Claims Resolution frequency and described The second Claims Resolution frequency determines target customer ID and target insurance kind type, comprising:

The corresponding claim times of each Customer ID and the corresponding information table quantity of insuring of the Customer ID are counted, pass through first Claims Resolution frequency formula is calculated, and the corresponding first Claims Resolution frequency of each Customer ID is obtained；Wherein, the first Claims Resolution frequency Secondary formula is

The corresponding claim times of each insurance kind type and the corresponding insurance kind information table quantity of the insurance kind type are counted, is passed through Second Claims Resolution frequency formula is calculated, and the corresponding second Claims Resolution frequency of each insurance kind type is obtained；Wherein, described second Claims Resolution frequency formula be

If the first Claims Resolution frequency is greater than the first preset threshold, the corresponding Customer ID of the first Claims Resolution frequency is determined as The target customer ID；

If the second Claims Resolution frequency is greater than the second preset threshold, the corresponding insurance kind type of the second Claims Resolution frequency is determined For the target insurance kind type.

3. risk model method for building up as described in claim 1, which is characterized in that the Claims Resolution information table further includes Claims Resolution note Record；

It is described that information table of insuring is inquired according to the target customer ID, client characteristics dimension is obtained, and according to the target insurance kind Type queries insurance kind information table obtains insurance kind characteristic dimension, comprising:

It insures information table according to target customer ID inquiry, obtains target customer corresponding with the target customer ID and believe Claims Resolution in the corresponding Claims Resolution information table of the target customer ID is recorded and is obscured with target customer's information by breath Matching, extracts the client characteristics dimension from target customer's information；

According to the target insurance kind type queries insurance kind information table, target insurance kind letter corresponding with the target insurance kind type is obtained Claims Resolution record in the corresponding Claims Resolution information table of the target insurance kind type is carried out fuzzy with the target insurance kind information by breath Match, the insurance kind characteristic dimension is extracted from the target insurance kind information.

4. risk model method for building up as described in claim 1, which is characterized in that it is described according to the client characteristics dimension and The insurance kind characteristic dimension, is modeled using decision Tree algorithms, obtains risk model, comprising:

Corresponding weight is respectively configured to the client characteristics dimension and the insurance kind characteristic dimension；

According to the client characteristics dimension and the corresponding weight of the insurance kind characteristic dimension, feature weight table is obtained；

The feature weight table is modeled using ID3 algorithm, obtains the risk model.

5. risk model method for building up as claimed in claim 4, which is characterized in that described to use ID3 algorithm to the feature Weight table is modeled, and the risk model is obtained, comprising:

The feature weight table is traversed, according to client characteristics dimension described in the feature weight table and the insurance kind feature dimensions Degree, is calculated by desired value formula, comentropy formula and information gain formula, obtains each client characteristics dimension The information gain of information gain and each insurance kind characteristic dimension；Wherein, the desired value formula isThe comentropy formula isThe information gain formula is Gain (x, t)=Info (t)-Info (x, t), p_iIt is the probability of i-th of client characteristics dimension or insurance kind characteristic dimension, Info (t) be client characteristics dimension or insurance kind characteristic dimension desired value, m is the quantity of client characteristics dimension or insurance kind characteristic dimension, t It is client characteristics dimension or insurance kind characteristic dimension, x is all client characteristics dimensions or the corresponding weighted average of insurance kind characteristic dimension Value, x_iIt is i-th of client characteristics dimension or the corresponding weight of insurance kind characteristic dimension, Info (x, t) is client characteristics dimension or danger The comentropy of kind characteristic dimension, Gain (x, t) is the information gain of client characteristics dimension or insurance kind characteristic dimension；

Choose the root section for having the client characteristics dimension or the insurance kind characteristic dimension of maximum information gain as decision tree Point, and using the client characteristics dimension or the corresponding weight of the insurance kind characteristic dimension as the weight of the root node；

The root node based on the decision tree repeats the information gain for obtaining each client characteristics dimension and every The step of information gain of one insurance kind characteristic dimension, successively determine the decision tree second layer to each node of the last layer, until The corresponding characteristic value of lowest level node is all the same client characteristics dimension or the insurance kind characteristic dimension under each branch, is obtained The risk model.

6. a kind of risk matching process characterized by comprising

Obtain information to be matched of insuring；

Feature extraction is carried out to the information to be matched of insuring, obtains target customer's characteristic value and target insurance kind characteristic value respectively；

The risk model obtained using the risk model method for building up in claim any one of 1-5, to the target visitor Family characteristic value and the target insurance kind characteristic value carry out risk matching, obtain risk assessment value；

7. a kind of risk model establishes device characterized by comprising

Data obtaining module, for obtaining at least one Claims Resolution information table, the corresponding Customer ID of each Claims Resolution information table with One insurance kind type；

Target information determining module, for counting the corresponding first Claims Resolution frequency of each Customer ID and each insurance kind class The corresponding second Claims Resolution frequency of type determines target customer ID and mesh according to the first Claims Resolution frequency and the second Claims Resolution frequency Mark insurance kind type；

Characteristic dimension obtains module, for insuring information table according to target customer ID inquiry, obtains client characteristics dimension, and According to the target insurance kind type queries insurance kind information table, insurance kind characteristic dimension is obtained；

Risk model establishes module, for being calculated using decision tree according to the client characteristics dimension and the insurance kind characteristic dimension Method is modeled, and risk model is obtained.

8. a kind of risk coalignment characterized by comprising

Data obtaining module, for obtaining information to be matched of insuring；

Characteristics extraction module obtains target customer's feature for carrying out feature extraction to the information to be matched of insuring respectively Dimension and target insurance kind dimension；

Risk matching module, the risk for being obtained using the risk model method for building up in any one of claim 1-5 Model carries out risk matching to target customer's characteristic dimension and the target insurance kind dimension, obtains risk assessment value；

High risk information object information determination module, if being greater than default value-at-risk for the risk assessment value, it is described to It is high risk information with information of insuring.

9. a kind of computer equipment, including memory, processor and storage can transport in memory and on the processor Capable computer program, which is characterized in that the processor realizes such as claim 1 to 5 times when executing the computer program The step of one risk model method for building up；Alternatively, the processor realizes such as right when executing the computer program It is required that the step of 6 risk matching process.

10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists In realizing the risk model method for building up as described in any one of claim 1 to 5 when the computer program is executed by processor Step；Alternatively, the computer program realizes the step of risk matching process as claimed in claim 6 when being executed by processor.