CN109685543A - Method, apparatus and computer equipment based on machine learning definite policy present - Google Patents
Method, apparatus and computer equipment based on machine learning definite policy present Download PDFInfo
- Publication number
- CN109685543A CN109685543A CN201811320261.6A CN201811320261A CN109685543A CN 109685543 A CN109685543 A CN 109685543A CN 201811320261 A CN201811320261 A CN 201811320261A CN 109685543 A CN109685543 A CN 109685543A
- Authority
- CN
- China
- Prior art keywords
- present
- client
- information
- value
- threshold value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000010801 machine learning Methods 0.000 title claims abstract description 29
- 238000003066 decision tree Methods 0.000 claims description 49
- 238000012549 training Methods 0.000 claims description 38
- 239000013598 vector Substances 0.000 claims description 22
- 238000004590 computer program Methods 0.000 claims description 13
- 238000005457 optimization Methods 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 8
- 241001269238 Data Species 0.000 claims description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000007477 logistic regression Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 230000006735 deficit Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000012417 linear regression Methods 0.000 description 4
- 238000005303 weighing Methods 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 230000005611 electricity Effects 0.000 description 3
- 238000012790 confirmation Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000000630 rising effect Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0207—Discounts or incentives, e.g. coupons or rebates
- G06Q30/0213—Consumer transaction fees
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Technology Law (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
This application involves field of artificial intelligence, disclose a kind of method, apparatus and computer equipment based on machine learning definite policy present, and wherein method includes: to obtain the policy information of the declaration form of client；According to the policy information, the corresponding present value threshold value of declaration form that can feed back to the client is calculated；Bonus information by the present value in bonus data library lower than the present of present value threshold value loads in the corresponding client of the client；At least one present of client's selection is received, and calculates the aggregate value of the corresponding value of each present of client's selection；Judge whether the aggregate value is less than or equal to the present value threshold value；If so, determining that at least one present of client's selection belongs to the present of the client.The application is calculated according to the information of customer information and the motor vehicle insurance business of purchase is suitble to give present corresponding to customer value, rationally pushes list of the present value within the scope of the declaration form of client to client.
Description
Technical field
This application involves field of artificial intelligence is arrived, especially relate to a kind of based on machine learning definite policy present
Method, apparatus and computer equipment.
Background technique
With electricity sell channel sales product by way of it is more and more common, in the sales process of motor vehicle insurance business
Start to be sold by the way of selling channel by electricity.The existing process that channel sales motor vehicle insurance business is sold by electricity
In, it generally can be improving client's attraction in a manner of Presents Giving, present giving system is typically all logical on existing line
The mode Presenting gifts arranged under line are crossed, are disconnected so as to cause giving for gift with filing statistics, later period backtracking property is poor；Present
It is all that insurance company purchases, if insurance company all shows present, it is possible to be that insurance expenses is less
Client select the higher present of purchase cost, will lead in this way company for the client declaration form lose.If present is pressed
It is placed according to the cost classification of buying, can have the case where client cannot but select when seeing the present liked again in this way, in this way
It can give client bad experience again.Therefore how to provide a kind of according to customer information and the information of the motor vehicle insurance business of purchase
The method for helping client to select declaration form present becomes urgent problem to be solved.
Summary of the invention
The main purpose of the application be a kind of automatic push are provided and assist client select declaration form present based on engineering
Practise the method, apparatus and computer equipment of definite policy present.
In order to achieve the above-mentioned object of the invention, the application proposes a kind of method based on machine learning definite policy present, packet
It includes:
The policy information of the declaration form of client is obtained, the policy information includes insurance products title, information of vehicles, premium letter
Breath；
According to the policy information, the corresponding present value threshold value of declaration form that can feed back to the client is calculated；
Bonus information by the present value in bonus data library lower than the present of present value threshold value is loaded in institute
It states in the corresponding client of client, the bonus data library is for storing bonus information and the corresponding present valence of bonus information
Value；
At least one present of client's selection is received, and calculate the corresponding value of each present of client's selection
Aggregate value；
Judge whether the aggregate value is less than or equal to the present value threshold value；
If so, determining that at least one present of client's selection belongs to the present of the client.
Further, described according to the policy information, calculate the corresponding present valence of declaration form that can feed back to the client
The step of being worth threshold value, comprising:
The policy information of client is input in the decision tree after the training of the policy information based on historic customer, the declaration form
Information includes that the history insurance of the client records regional information locating for information, motor vehicle insurance business, information of vehicles, premium letter
Breath；
Receive the first classification results of the decision tree output；
Present value threshold value corresponding with the declaration form is matched to according to first classification results.
Further, described according to the policy information, calculate the corresponding present valence of declaration form that can feed back to the client
Before the step of being worth threshold value, comprising:
Multiple sample datas in preset training set are input in preset decision tree, the training set includes multiple
The policy information of historic customer；
The testing classification of preset decision tree output is received as a result, comparing with preset sample classification result；
If comparing result is consistent, determine that the preset decision tree is the decision tree after training.
Further, described according to the policy information, calculate the step that can feed back to the present value threshold value of the client
After rapid, and in the bonus information by by the present value in bonus data library lower than the present of present value threshold value
Load is before the step in the corresponding client of the client, comprising:
The customer information for obtaining the client, the Logic Regression Models after being input to training, be not in danger coefficient, described
The coefficient that is not in danger indicates the probability that the client is not in danger；
The coefficient that is not in danger is worth threshold value multiplied by the present, updates the present value threshold value.
Further, the customer information for obtaining the client, the Logic Regression Models after being input to training obtain not
Be in danger coefficient the step of before, comprising:
The customer information of multiple clients being in danger is input to GBDT model, be not in danger vector；
The corresponding coefficient of each vector, the i.e. characteristic coefficient of Logic Regression Models are calculated by iteration optimization algorithms.
Further, present of the present value by bonus data library lower than the present of present value threshold value
Information loads the step in the corresponding client of the client, comprising:
Filter out target present of the present value less than or equal to present value threshold value；
The target present is ranked up according to the sequence that the present of the target present is worth from high to low；
The target present after sequence is loaded in the client.
Further, present of the present value by bonus data library lower than the present of present value threshold value
Information loads the step in the corresponding client of the client, comprising:
The gender for reading customer information gives present identical with the gender value in bonus data library lower than described
The bonus information that product are worth the present of threshold value loads in the corresponding client of the client.
The application also provides a kind of device based on machine learning definite policy present, comprising:
Module is obtained, the policy information of the declaration form for obtaining client, the policy information includes insurance products title, vehicle
Information, premium information；
First computing module can feed back to the declaration form of the client and corresponding give for calculating according to the policy information
Product are worth threshold value；
Loading module, for the present by the present value in bonus data library lower than the present of present value threshold value
Information loads in the corresponding client of the client, and the bonus data library is for storing bonus information and bonus information pair
The present value answered；
Second computing module for receiving at least one present of client's selection, and calculates client's selection
The aggregate value of the corresponding value of each present；
Judgment module, for judging whether the aggregate value is less than or equal to the present value threshold value；
Determination module determines client's choosing if being worth threshold value less than or equal to the present for the aggregate value
At least one present selected belongs to the present of the client.
The application also provides a kind of computer equipment, including memory and processor, and the memory is stored with computer
The step of program, the processor realizes any of the above-described the method when executing the computer program.
The application also provides a kind of computer readable storage medium, is stored thereon with computer program, the computer journey
The step of method described in any of the above embodiments is realized when sequence is executed by processor.
The method, apparatus and computer equipment based on machine learning definite policy present of the application, according to customer information
And the information of the motor vehicle insurance business of purchase calculates and is suitble to give present corresponding to customer value, rationally pushes present valence
It is worth the list within the scope of the declaration form of client to client, keeps the present given value not too high and make company deficit, also make to give
The present sent is worth service experience effect that will not be too low and bad to client.
Detailed description of the invention
Fig. 1 is the flow diagram of the method based on machine learning definite policy present of one embodiment of the application；
Fig. 2 is the structural schematic block diagram of the device based on machine learning definite policy present of one embodiment of the application；
Fig. 3 is the structural schematic block diagram of the device based on machine learning definite policy present of one embodiment of the application；
Fig. 4 is the structural schematic block diagram of the device based on machine learning definite policy present of one embodiment of the application；
Fig. 5 is that the structure of the loading module of the device based on machine learning definite policy present of one embodiment of the application is shown
Meaning block diagram；
Fig. 6 is that the structure of the loading module of the device based on machine learning definite policy present of one embodiment of the application is shown
Meaning block diagram；
Fig. 7 is the structural schematic block diagram of the computer equipment of one embodiment of the application.
The embodiments will be further described with reference to the accompanying drawings for realization, functional characteristics and the advantage of the application purpose.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Referring to Fig.1, the embodiment of the present application provides a kind of method based on machine learning definite policy present, comprising steps of
S1, obtain client declaration form policy information, the policy information includes insurance products title, information of vehicles, guarantor
Charge information；
S2, according to the policy information, calculate the corresponding present value threshold value of declaration form that can feed back to the client；
S3, the bonus information load by the present value in bonus data library lower than the present of present value threshold value exist
In the corresponding client of the client, the bonus data library is for storing bonus information and the corresponding present valence of bonus information
Value；
S4, at least one present for receiving client's selection, and calculate the corresponding valence of each present of client's selection
The aggregate value of value；
S5, judge whether the aggregate value is less than or equal to the present value threshold value；
S6, if so, determining that at least one present of client selection belongs to the present of the client.
As described in above-mentioned steps S1, client can register an account, then client when buying insurance by the network platform
The network platform is accessed by client logon account and fills in insurance information, generates declaration form.Client generates one by client
After declaration form, server gets the policy information of the client.When client insures on the client, need to fill in insurer's letter
Breath-i.e. customer information, further includes the information-i.e. information of vehicles of insurance subject, also mode-i.e. insurance products the title of insurance,
There are also the quantity-i.e. of premium premium informations.After client fills in these information by client, confirmation is submitted, that is, completes to insure.
After client has submitted policy information by client, server receives the policy information that client is filled in.Client by mobile phone,
Hardware access server that computer etc. can surf the Internet simultaneously carries out filling in policy information after logging in oneself account again.
As described in above-mentioned steps S2, each present is to give after insurance company voluntarily buys to client, therefore present
It is containing certain cost；The client more to profit, it is corresponding to pay more costs, therefore the value of present
Can be higher, but can not be too high, too Gao Huishigai insurance company violates relevant legal provisions or leads to company pair
The declaration form is in the state of loss.Thus one present of setting is worth threshold value, it can feeds back to the highest price of the present of client
Value.After server gets policy information, according to preset computation rule, the corresponding visitor of the declaration form can be fed back to by being calculated
The present of the gift value at family is worth threshold value, wherein present value threshold value is calculated for the policy information of each declaration form
Come.Same insurer generates different policy informations because buying different insurances, and corresponding present value threshold value is not phase
With.
As described in above-mentioned steps S3, bonus data library is that a whole for being stored with the present that can be given in server is believed
The database of breath, bonus information include present type, present specification, present value, present image etc..Server gets above-mentioned
After present is worth threshold value, the present value of all presents in bonus data library is read, filters out present value lower than present valence
It is worth the corresponding bonus information of present of threshold value, then all loads the bonus information screened on the client, Ke Hutong
The present that can be given can be checked by crossing client.
As described in above-mentioned steps S4, client selects oneself favorite present from the present that can be given, and server receives
After the present selected to client, the present value for the present chosen is added, present aggregate value is obtained.
As described in above-mentioned steps S5, server calculate client selection present aggregate value after, then by this and with give
Product value threshold value is compared, and sees whether be more than that present is worth threshold value, if it does, explanation does not meet corporate policy.
As described in above-mentioned steps S6, it is worth threshold value when aggregate value is less than or equal to present, further relates to client's selection
Present is giving in range in company, illustrates that the present of client's selection does not violate relevant regulation, is possible.Service
Device determines that the present of client's selection is the present for belonging to client.
In one embodiment, above-mentioned according to the policy information, the declaration form that calculating can feed back to the client is corresponding
The step S2 of present value threshold value, comprising:
S21, the policy information of client is input in the decision tree after the policy information based on historic customer is trained, it is described
Policy information further include the history insurance record information of the client, regional information, information of vehicles locating for motor vehicle insurance business,
Premium information；
S22, the first classification results for receiving the decision tree output；
S23, present value threshold value corresponding with the declaration form is matched to according to first classification results.
In the present embodiment, server reads policy information, and server insures record information, vapour according to the former years of historic customer
Regional information locating for vehicle insurance business, client information of vehicles obtain the present total value that can be given, by by client's
The former years insurance of regional information locating for information of vehicles, motor vehicle insurance business, client, which records, is input to preset first decision tree
In classified to obtain the first classification results, server receives the first classification results of decision tree output, then according to first point
Class matches to obtain the present value threshold value of the present that can be given corresponding with above-mentioned declaration form.
In one embodiment, above-mentioned according to the policy information, the declaration form that calculating can feed back to the client is corresponding
Present is worth before the step S2 of threshold value, comprising:
S201, multiple sample datas in preset training set are input in preset decision tree, the training set packet
Include the policy information of multiple historic customers；
S202, the class test of preset decision tree output is received as a result, comparing with preset sample classification result；
If S203, comparing result are consistent, determine that the preset decision tree is the decision tree after training.
In the present embodiment, by the former years insurance record of the historic customer in the sample data of test set, motor vehicle insurance business
The information of vehicles of locating regional information, historic customer is input to preset decision tree and is classified to obtain classification prediction result,
By by above-mentioned classification prediction result and the history of the client of input insurance record, regional information locating for motor vehicle insurance business,
Sample classification result corresponding to the information of vehicles of client compares, and sees whether the two is consistent, if the two result is consistent, verifying
Pass through, then illustrates that preset decision tree training is completed.For different sample classifications as a result, it is desirable to which explanation, meeting is according to visitor
The former years insurance record at family, regional information locating for motor vehicle insurance business, the information of vehicles of client these three dimensions information set
Different nodal informations is set to classify.Firstly, for client former years insure record, it includes specifying information be whether have
Buy the record of our company's motor vehicle insurance business, can according to having purchase, be arranged two different nodes without buying, for
Above-mentioned different class node also continues the information according to this dimension of regional information locating for motor vehicle insurance business point
Class.For this dimensional information of regional information locating for motor vehicle insurance business, by according to locating for motor vehicle insurance business differently
Different classifications node is arranged to be arranged the different city such as different nodes, such as Shenzhen, Dongguan and Huizhou in domain information, right
Also the information according to information of vehicles this dimension for obtaining client is carried out continuing to classify in above-mentioned sample classification result.For visitor
The information of vehicles at family, the information of vehicles of above-mentioned client are specially that client is buying the motor vehicle insurance business institute putting expense, for
Different nodal informations can be arranged in the purchase motor vehicle insurance business institute putting expense in client, wherein the above-mentioned expense according to investment
Setting 0-5000 member, 5000-10000 member, 10000-15000 member, 15000-20000 member and 20000 yuan are with first-class different
Class node.
In one embodiment, above-mentioned according to the policy information, calculate the present value threshold that can feed back to the client
After the step S2 of value, and in the present that the present value in bonus data library will be lower than to present value threshold value
Bonus information loads before the step S3 in the corresponding client of the client, comprising:
S211, the customer information for obtaining the client, the Logic Regression Models after being input to training, not being in danger is
Number, the coefficient that is not in danger indicate the probability that the client is not in danger；
S212, the coefficient that is not in danger is worth threshold value multiplied by the present, updates the present value threshold value.
In the present embodiment, customer information includes gender, age, the driver's license time limit, educational background of client etc., because of the driving of client
It is accustomed to having certain degree of association with the customer information of client and proprietary information, for example older people's driving is more steady, it is right
The probability being in danger answered is smaller；The high relatively common quality of people of educational background is higher, corresponding to have better driving habit, then is in danger
Probability it is more relatively small.Therefore, customer information can be integrated to calculate the probability that is not in danger of client.Client is filling in declaration form
When information, above-mentioned customer information is the content having to fill out.Server obtains the customer information in declaration form.Logic Regression Models
Using being mainly used for probability expression, it is fast that the advantages of model, is to solve for speed, using convenient.In the training logistic regression mould
When type, first inputs multiple customer informations for generating Claims Resolution and calculate the classification pair of customer information then according to the classification of customer information
The weighing factor of Claims Resolution is not generated.Logic Regression Models are based on these weighing factors again, calculate the client according to customer data
The probability of Claims Resolution is not generated after insurance products on order.Customer information is input to the Logic Regression Models after training
Calculate the probability that is not in danger of the corresponding customer data of the client.The coefficient that is not in danger of client is higher, then it represents that client more will not
It is in danger, the probability that corresponding insurance company compensates is lower, and the money earned is more, can be higher with the value of the present of feedback.
Therefore, it is necessary to which present value threshold value to be updated, the coefficient that will not be in danger is worth threshold value multiplied by present, to update present value threshold
Value.
In one embodiment, the customer information of the above-mentioned acquisition client, the Logic Regression Models after being input to training,
It is not in danger before the step S211 of coefficient, comprising:
S204, the customer information of multiple clients not being in danger is input to GBDT model, be not in danger vector；
S205, the corresponding coefficient of each vector, the i.e. characteristic coefficient of Logic Regression Models are calculated by iteration optimization algorithms.
As described in above-mentioned steps S204, the client not being in danger refers to the insurance products for the company of having purchased and in period insured
The client that the insurance thing such as friction, collision is settled a claim does not occur.Customer information include age of client, gender, automobile model,
Vehicle price, the driver's license age of client, automobile age etc..GBDT model is called gradient boosted tree (Gradient Boosting
Decison Tree) model.According to the basic data of the client got, and the GBDT model of different levels is set.Gradient mentions
Rising tree has an at least decision tree, each tree to have multiple leaf nodes.According to the type of data and dimension, different numbers are set
One customer information is input in decision tree by the leaf node of amount, exports a vector, the i.e. pre- direction finding that is not in danger of client
Amount.In a simple embodiment, using the gender in user information as the vector that is not in danger, then there are two leaf node, settings for tool
First leaf node is gender male, and it is not male, i.e. women that second leaf node, which is arranged, as gender.By the data of a client
It is input in GBDT model, if male client, then falls on first leaf node, obtained predicted vector of not being in danger is
(1,0).In a particular embodiment, corresponding according to multiple information dimensions, the dimension of predicted vector can be more.Do not go out multiple
The user information of the client of danger is input in GBDT model, obtains multiple vectors that are not in danger.
As described in above-mentioned steps S205, logistic regression is a kind of generalized linear regression, is added on the basis of linear regression
Sigmoid function is entered and has carried out Nonlinear Mapping, successive value can be mapped on 0 and 1 by this function, a wide range of numerical value pressure
It is reduced within the scope of this, the influence of the variable especially to stand out can be eliminated, that is, eliminate the exceptional value of data.Logistic regression because become
Amount can be two classification be also possible to it is polytypic, in practice commonly is exactly two classify logistic regressions.It is calculated by iteration
Method optimization algorithm determines the characteristic coefficient of the Logic Regression Models.Iteration optimization algorithms can be L-BFGS algorithm and be also possible to
SGD algorithm.L-BFGS algorithm be it is a kind of solve without the constraint common method of linear optimization problem, which has more perfect part
Convergence theory, it is advantageous on large data sets.
In one embodiment, present of the above-mentioned present value by bonus data library lower than present value threshold value
Bonus information load step S3 in the corresponding client of the client, comprising:
S31, target present of the present value less than or equal to present value threshold value is filtered out；
S32, the target present is ranked up according to the sequence that the present of the target present is worth from high to low；
S33, the target present after sequence is loaded in the client.
In the present embodiment, bonus information of the server by the present value in bonus data library lower than present value threshold value is sieved
It elects, is then worth according to present and is ranked up according to sequence from high to low, it is however generally that, present value is higher to give
Product are more liked by client, and present is ranked up according to the height that present is worth, then loads on the client, is convenient for
Client selects better, more good present.
In one embodiment, present of the above-mentioned present value by bonus data library lower than present value threshold value
Bonus information load step S3 in the corresponding client of the client, comprising:
Present identical with gender value in bonus data library is lower than institute by S34, the gender for reading customer information
The bonus information for stating the present of present value threshold value loads in the corresponding client of the client.
In the present embodiment, after server is screened according to present value, obtain to give to the present of client, so
The bonus information for reading the present that can be given afterwards, reads out the gender label to present.Then the client in customer information is read
Then gender loads bonus information corresponding with the consistent gender label of client gender in client.General male client's happiness
The more practical presents such as joyous handset bracket, direction indicators cover, general women client, which likes male earner, pillow etc., can decorate automobile
Present.Corresponding present is loaded in client according to the gender of client, the favorite present of client can be only loaded, more improve
Customer experience.
In conclusion the application's determines the method for obtaining declaration form present based on machine, according to customer information and purchase
The information of motor vehicle insurance business calculate and be suitble to give present corresponding to customer value, rationally push present value in client
Declaration form within the scope of list to client, keep the present given value not too high and make company deficit, also make the present given
It is worth service experience effect that will not be too low and bad to client.
Referring to Fig. 2, a kind of device for determining based on machine and obtaining declaration form present is also provided in the embodiment of the present application, comprising:
Obtain module 1, the policy information of the declaration form for obtaining client, the policy information include insurance products title,
Information of vehicles, premium information；
First computing module 2 can feed back to the declaration form of the client and corresponding give for calculating according to the policy information
Product are worth threshold value；
Loading module 3, for the present in bonus data library to be worth to giving for the present lower than present value threshold value
Product information loads in the corresponding client of the client, and the bonus data library is for storing bonus information and bonus information
Corresponding present value；
Second computing module 4 for receiving at least one present of client's selection, and calculates client's selection
The aggregate value of the corresponding value of each present；
Judgment module 5, for judging whether the aggregate value is less than or equal to the present value threshold value；
Determination module 6 determines client's choosing if being worth threshold value less than or equal to the present for the aggregate value
At least one present selected belongs to the present of the client.
In the present embodiment, client can register an account, then client passes through visitor when buying insurance by the network platform
The family end logon account access network platform fills in insurance information, generates declaration form.After client generates a declaration form by client,
Obtain the policy information that module 1 gets the client.When client insures on the client, need to fill in insurer's information-
That is customer information further includes the information-i.e. information of vehicles of insurance subject, and there are also the mode-i.e. insurance products titles of insurance, also
The quantity-i.e. of premium premium information.After client fills in these information by client, confirmation is submitted, that is, completes to insure.Client
After having submitted policy information by client, server receives the policy information that client is filled in.Client passes through mobile phone, computer
It carries out filling in policy information again etc. the hardware access server that can be surfed the Internet and after logging in oneself account.
Each present is to be given after insurance company voluntarily buys to client, therefore present is containing certain cost
's；The client more to profit, it is corresponding to pay more costs, therefore the value of present can be higher, but
Cannot be too high, too Gao Huishigai insurance company violates relevant legal provisions or company is caused to be in the shape lost to the declaration form
State.Thus one present of setting is worth threshold value, it can feeds back to the maximum value of the present of client.Module 1 is obtained to get
After policy information, for the first computing module 2 according to preset computation rule, the corresponding client of the declaration form can be fed back to by being calculated
The present of gift value be worth threshold value, wherein present value threshold value is calculated for the policy information of each declaration form
's.Same insurer generates different policy informations because buying different insurances, and corresponding present value threshold value is not identical
's.
Bonus data library is the database for all information that one in server is stored with the present that can be given, present letter
Breath includes present type, present specification, present value, present image etc..After server gets above-mentioned present value threshold value, read
It takes the present of all presents in bonus data library to be worth, it is corresponding lower than the present of present value threshold value to filter out present value
Bonus information, then by the bonus information screened, all on the client, client is loading module 3 by client for load
It can check the present that can be given.
Client selects oneself favorite present from the present that can be given, and the second computing module 4 receives client's selection
Present after, the present of the present chosen value is added, present aggregate value is obtained.
Server calculate client selection present aggregate value after, then judgment module 5 should and with present value threshold
Value is compared, and sees whether be more than that present is worth threshold value, if it does, explanation does not meet corporate policy.
When aggregate value is worth threshold value less than or equal to present, the present for further relating to client's selection is giving in company
It send in range, illustrates that the present of client's selection does not violate relevant regulation, be possible.Determination module 6 decides that client selects
The present selected is the present for belonging to client.
Referring to Fig. 3, in one embodiment, above-mentioned first computing module 2 includes:
Input unit 21, for the policy information of client to be input to determining after the policy information based on historic customer is trained
In plan tree, the policy information further includes that the history insurance of the client records region letter locating for information, motor vehicle insurance business
Breath, information of vehicles, premium information；
Receiving unit 22, for receiving the first classification results of the decision tree output；
Matching unit 23, for being matched to present value threshold corresponding with the declaration form according to first classification results
Value.
In the present embodiment, server reads policy information, and server insures record information, vapour according to the former years of historic customer
Regional information locating for vehicle insurance business, client information of vehicles obtain the present total value that can be given, input unit 21 will
Regional information locating for the information of vehicles of client, motor vehicle insurance business, client former years insurance record be input to preset first
Classified to obtain the first classification results in decision tree, receiving unit 22 receives the first classification results of decision tree output, then
Matching unit 23 is worth threshold value according to the present that the first classification and matching obtains the present that can be given corresponding with above-mentioned declaration form.
In one embodiment, the above-mentioned device based on machine learning definite policy present, further includes:
First input module 201, for multiple sample datas in preset training set to be input to preset decision tree
In, the training set includes the policy information of multiple historic customers；
Contrast module 202, for receiving the class test of preset decision tree output as a result, with preset sample classification knot
Fruit compares；
Determining module 203 determines that the preset decision tree is the decision tree after training if consistent for comparing result.
In the present embodiment, the former years of the historic customer in the sample data of test set are insured note by the first input module 201
It records, the information of vehicles of regional information locating for motor vehicle insurance business, historic customer is input to preset decision tree and classify
To classification prediction result, contrast module 202 protects above-mentioned classification prediction result and the history of the client of input insurance record, automobile
Regional information locating for dangerous business, client information of vehicles corresponding to sample classification result compare, both see whether one
It causes, if the two result is consistent, is verified, then illustrate that preset decision tree training is completed, determining module 203 determines the decision tree
For the decision tree after training.For different sample classifications as a result, it is desirable to which explanation, can insure note according to the former years of client
Record, regional information locating for motor vehicle insurance business, the information of vehicles of client these three dimensions information different node letters is set
Breath is classified.Firstly, for client former years insure record, it includes specifying information be whether have purchase our company's automobile
The record of insurance business, can be according to having purchase, two different nodes being arranged without purchase, for above-mentioned different classification
Information according to this dimension of regional information locating for motor vehicle insurance business also is continued to classify by node.Automobile is protected
This dimensional information of regional information locating for dangerous business will be arranged not according to different geographical information locating for motor vehicle insurance business
With the different city such as node, such as Shenzhen, Dongguan and Huizhou different classifications node is set, for above-mentioned sample classification
As a result also the information according to information of vehicles this dimension for obtaining client is carried out continuing to classify.For the information of vehicles of client,
The information of vehicles of above-mentioned client is specially that client is buying the motor vehicle insurance business institute putting expense, is buying the vapour for client
Different nodal informations can be arranged in vehicle insurance business institute putting expense, wherein it is above-mentioned according to the expense of investment be arranged 0-5000 member,
5000-10000 member, 10000-15000 member, 15000-20000 member and 20000 yuan are with first-class different class node.
Referring to Fig. 4, in one embodiment, the above-mentioned device based on machine learning definite policy present, further includes:
Coefficient module 211, for obtaining the customer information of the client, the Logic Regression Models after being input to training are obtained
To the coefficient that is not in danger, the coefficient that is not in danger indicates the probability that the client is not in danger；
Update module 212 updates the present value for the coefficient that is not in danger to be worth threshold value multiplied by the present
Threshold value.
In the present embodiment, customer information includes gender, age, the driver's license time limit, educational background of client etc., because of the driving of client
It is accustomed to having certain degree of association with the customer information of client and proprietary information, for example older people's driving is more steady, it is right
The probability being in danger answered is smaller；The high relatively common quality of people of educational background is higher, corresponding to have better driving habit, then is in danger
Probability it is more relatively small.Therefore, customer information can be integrated to calculate the probability that is not in danger of client.Client is filling in declaration form
When information, above-mentioned customer information is the content having to fill out.Coefficient module 211 obtains the customer information in declaration form.Logistic regression
The application of model is mainly used for probability expression, and it is fast that the advantages of model is to solve for speed, using convenient.In the training logic
When regression model, first inputs multiple customer informations for generating Claims Resolution and calculate customer information then according to the classification of customer information
Classification is to the weighing factor for not generating Claims Resolution.Logic Regression Models are based on these weighing factors again, are calculated according to customer data
The probability of Claims Resolution is not generated after insurance products on the customer order.After customer information is input to training by coefficient module 211
Logic Regression Models can calculate the probability that is not in danger of the corresponding customer data of the client.The coefficient that is not in danger of client is higher,
Then indicate that client will not more be in danger, the probability that corresponding insurance company compensates is lower, and the money earned is more, can giving with feedback
The value of product is higher.Therefore, it is necessary to update modules 212 to be updated present value threshold value, and the coefficient that will not be in danger is multiplied by giving
Product are worth threshold value, to update present value threshold value.
In one embodiment, the above-mentioned device based on machine learning definite policy present, further includes:
Second input module 204 obtains not for the customer information of multiple clients not being in danger to be input to GBDT model
Be in danger vector；
Design factor module 205, for calculating the corresponding coefficient of each vector, i.e. logistic regression mould by iteration optimization algorithms
The characteristic coefficient of type.
In the present embodiment, the client not being in danger refers to the insurance products for the company of having purchased and does not rub in period insured
The client that the insurance thing such as wiping, collision is settled a claim.Customer information includes the age of client, gender, automobile model, automobile valence
Lattice, the driver's license age of client, automobile age etc..GBDT model is called gradient boosted tree (Gradient Boosting
Decison Tree) model.According to the basic data of the client got, and the GBDT model of different levels is set.Gradient mentions
Rising tree has an at least decision tree, each tree to have multiple leaf nodes.According to the type of data and dimension, different numbers are set
One customer information is input in decision tree by the leaf node of amount, exports a vector, the i.e. pre- direction finding that is not in danger of client
Amount.In a simple embodiment, using the gender in user information as the vector that is not in danger, then there are two leaf node, settings for tool
First leaf node is gender male, and it is not male, i.e. women that second leaf node, which is arranged, as gender.By the data of a client
It is input in GBDT model, if male client, then falls on first leaf node, obtained predicted vector of not being in danger is
(1,0).In a particular embodiment, corresponding according to multiple information dimensions, the dimension of predicted vector can be more.Second input mould
The user information of multiple clients not being in danger is input in GBDT model by block 204, obtains multiple vectors that are not in danger.
Logistic regression is a kind of generalized linear regression, is that joined the progress of Sigmoid function on the basis of linear regression
Successive value can be mapped on 0 and 1 by Nonlinear Mapping, this function, can be with a wide range of magnitude compression within the scope of this
The influence of the variable especially to stand out is eliminated, that is, eliminates the exceptional value of data.The dependent variable of logistic regression can be two classification
Can be it is polytypic, in practice commonly is exactly two classify logistic regressions.Design factor module 205 is excellent by iterative algorithm
Change the characteristic coefficient that algorithm determines the Logic Regression Models.Iteration optimization algorithms can be L-BFGS algorithm and be also possible to SGD
Algorithm.L-BFGS algorithm be it is a kind of solve without the constraint common method of linear optimization problem, which has more perfect part to receive
Theory is held back, it is advantageous on large data sets.
Referring to Fig. 5, in one embodiment, above-mentioned loading module 3 includes:
Screening unit 31, for filtering out target present of the present value less than or equal to present value threshold value；
Sequencing unit 32, for the target present to be worth sequence from high to low according to the present of the target present
It is ranked up；
First loading unit 33 is loaded for the target present after sorting in the client.
In the present embodiment, present of the screening unit 31 by the present value in bonus data library lower than present value threshold value is believed
Breath screens, and then sequencing unit 32 is worth according to present and is ranked up according to sequence from high to low, it is however generally that, present
It is worth higher present, is more liked by client, present is ranked up according to the height that present is worth, then the first load
Unit 33 loads on the client, selects better, more good present convenient for client.
Referring to Fig. 6, in one embodiment, above-mentioned loading module 3 includes:
Second loading unit 34 will be identical as the gender in bonus data library for reading the gender of customer information
Present value lower than the present value threshold value present bonus information load in the corresponding client of the client.
In the present embodiment, after server is screened according to present value, obtain to give to the present of client, so
The bonus information for reading the present that can be given afterwards, reads out the gender label to present.Then the client in customer information is read
Gender, then the second loading unit 34 loads bonus information corresponding with the consistent gender label of client gender in client.
General male client likes the more practical present such as handset bracket, direction indicators cover, and general women client likes male earner, pillow etc.
The present of automobile can be decorated.Corresponding present is loaded in client according to the gender of client, can only load client be liked
Present, more raising customer experience.
In conclusion the device based on machine learning definite policy present of the application, according to customer information and purchase
The information of motor vehicle insurance business calculate and be suitble to give present corresponding to customer value, rationally push present value in client
Declaration form within the scope of list to client, keep the present given value not too high and make company deficit, also make the present given
It is worth service experience effect that will not be too low and bad to client.
Referring to Fig. 7, a kind of computer equipment is also provided in the embodiment of the present application, which can be server,
Its internal structure can be as shown in Figure 7.The computer equipment includes processor, the memory, network connected by system bus
Interface and database.Wherein, the processor of the Computer Design is for providing calculating and control ability.The computer equipment is deposited
Reservoir includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program
And database.The internal memory provides environment for the operation of operating system and computer program in non-volatile memory medium.It should
The database of computer equipment is for storing the data such as customer information.The network interface of the computer equipment is used for and external end
End passes through network connection communication.To realize that a kind of machine learning definite policy that is based on is given when the computer program is executed by processor
The method of product.
Above-mentioned processor executes the step of above-mentioned method based on machine learning definite policy present: obtaining the declaration form of client
Policy information, the policy information includes insurance products title, information of vehicles, premium information；According to the policy information, meter
Calculate the corresponding present value threshold value of declaration form that can feed back to the client；Present value in bonus data library is given lower than described
The bonus information that product are worth the present of threshold value loads in the corresponding client of the client, and the bonus data library is for storing
Bonus information and the corresponding present value of bonus information；At least one present of client's selection is received, and described in calculating
The aggregate value of the corresponding value of each present of client's selection；Judge whether the aggregate value is less than or equal to the present valence
It is worth threshold value；If so, determining that at least one present of client's selection belongs to the present of the client.
In one embodiment, above-mentioned processor is executed according to the policy information, and calculating can feed back to the client's
The step of corresponding present of declaration form is worth threshold value, comprising: the policy information of client is input to the letter of the declaration form based on historic customer
In decision tree after breath training, the policy information includes history insurance the record information, motor vehicle insurance business institute of the client
The regional information at place, information of vehicles, premium information；Receive the first classification results of the decision tree output；According to described first
Classification results are matched to present value threshold value corresponding with the declaration form.
In one embodiment, above-mentioned processor is executed according to the policy information, and calculating can feed back to the client's
Before the step of corresponding present of declaration form is worth threshold value, comprising: be input to multiple sample datas in preset training set pre-
If decision tree in, the training set includes the policy information of multiple historic customers；Receive the test of preset decision tree output
Classification results are compared with preset sample classification result；If comparing result is consistent, determine the preset decision tree for instruction
Decision tree after white silk.
In one embodiment, above-mentioned processor is executed according to the policy information, and calculating can feed back to the client's
Present was worth after the step of threshold value, and the present value in bonus data library will be worth threshold value lower than the present described
Present bonus information load before the step in the corresponding client of the client, comprising: obtain the visitor of the client
Family information, the Logic Regression Models after being input to training, be not in danger coefficient, and the coefficient that is not in danger indicates that the client does not go out
The probability of danger；The coefficient that is not in danger is worth threshold value multiplied by the present, updates the present value threshold value.
In one embodiment, above-mentioned processor executes the customer information for obtaining the client, patrolling after being input to training
Volume regression model, be not in danger coefficient the step of before, comprising: the customer information of multiple clients being in danger is input to
GBDT model, be not in danger vector；By iteration optimization algorithms calculate the corresponding coefficient of each vector, i.e. Logic Regression Models
Characteristic coefficient.
In one embodiment, above-mentioned processor, which is executed, is worth the present value in bonus data library lower than the present
The bonus information of the present of threshold value loads the step in the corresponding client of the client, comprising: it is low to filter out present value
In or equal to present value threshold value target present；By the target present according to the target present present value from
High to Low sequence is ranked up；The target present after sequence is loaded in the client.
In one embodiment, above-mentioned processor, which is executed, is worth the present value in bonus data library lower than the present
The bonus information of the present of threshold value loads the step in the corresponding client of the client, comprising: reads the property of customer information
Not, the present by present identical with the gender value in bonus data library lower than the present of present value threshold value is believed
Breath load is in the corresponding client of the client.
In conclusion information meter of the computer equipment of the application according to customer information and the motor vehicle insurance business of purchase
Calculating is suitble to give present corresponding to customer value, rationally pushes list of the present value within the scope of the declaration form of client to visitor
Family keeps the present given value not too high and make company deficit, also make the present given value will not it is too low and not to client
Good service experience effect.
It will be understood by those skilled in the art that structure shown in Fig. 7, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme.
One embodiment of the application also provides a kind of computer readable storage medium, is stored thereon with computer program, calculates
Machine program realizes a kind of method based on machine learning definite policy present when being executed by processor, specifically: obtain client's
The policy information of declaration form, the policy information include insurance products title, information of vehicles, premium information；Believed according to the declaration form
Breath calculates the corresponding present value threshold value of declaration form that can feed back to the client；Present value in bonus data library is lower than
The bonus information of the present of the present value threshold value loads in the corresponding client of the client, and the bonus data library is used
In storage bonus information and the corresponding present value of bonus information；At least one present of client's selection is received, and is counted
Calculate the aggregate value of the corresponding value of each present of client's selection；It is described to judge whether the aggregate value is less than or equal to
Present is worth threshold value；If so, determining that at least one present of client's selection belongs to the present of the client.
In one embodiment, above-mentioned processor is executed according to the policy information, and calculating can feed back to the client's
The step of corresponding present of declaration form is worth threshold value, comprising: the policy information of client is input to the letter of the declaration form based on historic customer
In decision tree after breath training, the policy information includes history insurance the record information, motor vehicle insurance business institute of the client
The regional information at place, information of vehicles, premium information；Receive the first classification results of the decision tree output；According to described first
Classification results are matched to present value threshold value corresponding with the declaration form.
In one embodiment, above-mentioned processor is executed according to the policy information, and calculating can feed back to the client's
Before the step of corresponding present of declaration form is worth threshold value, comprising: be input to multiple sample datas in preset training set pre-
If decision tree in, the training set includes the policy information of multiple historic customers；Receive the test of preset decision tree output
Classification results are compared with preset sample classification result；If comparing result is consistent, determine the preset decision tree for instruction
Decision tree after white silk.
In one embodiment, above-mentioned processor is executed according to the policy information, and calculating can feed back to the client's
Present was worth after the step of threshold value, and the present value in bonus data library will be worth threshold value lower than the present described
Present bonus information load before the step in the corresponding client of the client, comprising: obtain the visitor of the client
Family information, the Logic Regression Models after being input to training, be not in danger coefficient, and the coefficient that is not in danger indicates that the client does not go out
The probability of danger；The coefficient that is not in danger is worth threshold value multiplied by the present, updates the present value threshold value.
In one embodiment, above-mentioned processor executes the customer information for obtaining the client, patrolling after being input to training
Volume regression model, be not in danger coefficient the step of before, comprising: the customer information of multiple clients being in danger is input to
GBDT model, be not in danger vector；By iteration optimization algorithms calculate the corresponding coefficient of each vector, i.e. Logic Regression Models
Characteristic coefficient.
In one embodiment, above-mentioned processor, which is executed, is worth the present value in bonus data library lower than the present
The bonus information of the present of threshold value loads the step in the corresponding client of the client, comprising: it is low to filter out present value
In or equal to present value threshold value target present；By the target present according to the target present present value from
High to Low sequence is ranked up；The target present after sequence is loaded in the client.
In one embodiment, above-mentioned processor, which is executed, is worth the present value in bonus data library lower than the present
The bonus information of the present of threshold value loads the step in the corresponding client of the client, comprising: reads the property of customer information
Not, the present by present identical with the gender value in bonus data library lower than the present of present value threshold value is believed
Breath load is in the corresponding client of the client.
In conclusion the computer readable storage medium of the application, according to customer information and the car insurance industry of purchase
The information of business, which calculates, is suitble to give present corresponding to customer value, rationally pushes present value within the scope of the declaration form of client
List to client, keep the present given value not too high and make company deficit, also make the present given value will not be too low
And the service experience effect bad to client.
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
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
Any reference used in provided herein and embodiment to memory, storage, database or other media,
Including 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 speed are according to rate SDRAM (SSRSDRAM), 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 should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, device, article or the method that include a series of elements not only include those elements, and
And further include other elements that are not explicitly listed, or further include for this process, device, article or method institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do
There is also other identical elements in the process, device of element, article or method.
The foregoing is merely preferred embodiment of the present application, are not intended to limit the scope of the patents of the application, all utilizations
Equivalent structure or equivalent flow shift made by present specification and accompanying drawing content is applied directly or indirectly in other correlations
Technical field, similarly include in the scope of patent protection of the application.
Claims (10)
1. a kind of method based on machine learning definite policy present characterized by comprising
The policy information of the declaration form of client is obtained, the policy information includes insurance products title, information of vehicles, premium information；
According to the policy information, the corresponding present value threshold value of declaration form that can feed back to the client is calculated；
Bonus information by the present value in bonus data library lower than the present of present value threshold value is loaded in the visitor
In the corresponding client in family, the bonus data library is for storing bonus information and the corresponding present value of bonus information；
At least one present of client's selection is received, and calculates the value of the corresponding value of each present of client's selection
Summation；
Judge whether the aggregate value is less than or equal to the present value threshold value；
If so, determining that at least one present of client's selection belongs to the present of the client.
2. the method as described in claim 1 based on machine learning definite policy present, which is characterized in that described according to
The step of policy information, the corresponding present of declaration form that calculating can feed back to the client is worth threshold value, comprising:
The policy information of client is input in the decision tree after the training of the policy information based on historic customer, the policy information
Regional information, information of vehicles, premium information locating for history insurance record information, motor vehicle insurance business including the client；
Receive the first classification results of the decision tree output；
Present value threshold value corresponding with the declaration form is matched to according to first classification results.
3. the method as claimed in claim 2 based on machine learning definite policy present, which is characterized in that described according to
Before the step of policy information, the corresponding present of declaration form that calculating can feed back to the client is worth threshold value, comprising:
Multiple sample datas in preset training set are input in preset decision tree, the training set includes multiple history
The policy information of client；
The testing classification of preset decision tree output is received as a result, comparing with preset sample classification result；
If comparing result is consistent, determine that the preset decision tree is the decision tree after training.
4. the method as described in claim 1 based on machine learning definite policy present, which is characterized in that described according to
Policy information, calculate can feed back to the client present value threshold value the step of after, and it is described will be by bonus data library
In present value lower than the present value threshold value present bonus information load in the corresponding client of the client
The step of before, comprising:
The customer information for obtaining the client, the Logic Regression Models after being input to training, be not in danger coefficient, described not go out
Dangerous coefficient indicates the probability that the client is not in danger；
The coefficient that is not in danger is worth threshold value multiplied by the present, updates the present value threshold value.
5. the method as claimed in claim 4 based on machine learning definite policy present, which is characterized in that described in the acquisition
The customer information of client, be input to training after Logic Regression Models, be not in danger coefficient the step of before, comprising:
The customer information of multiple clients being in danger is input to GBDT model, be not in danger vector；
The corresponding coefficient of each vector, the i.e. characteristic coefficient of Logic Regression Models are calculated by iteration optimization algorithms.
6. the method as described in claim 1 based on machine learning definite policy present, which is characterized in that described by present number
Bonus information according to the present value in library lower than the present of present value threshold value is loaded in the corresponding client of the client
Step on end, comprising:
Filter out target present of the present value less than or equal to present value threshold value；
The target present is ranked up according to the sequence that the present of the target present is worth from high to low；
The target present after sequence is loaded in the client.
7. the method as described in claim 1 based on machine learning definite policy present, which is characterized in that described by present number
Bonus information according to the present value in library lower than the present of present value threshold value is loaded in the corresponding client of the client
Step on end, comprising:
Present identical with gender value in bonus data library is lower than the present valence by the gender for reading customer information
The bonus information for being worth the present of threshold value loads in the corresponding client of the client.
8. a kind of device based on machine learning definite policy present characterized by comprising
Module, the policy information of the declaration form for obtaining client are obtained, the policy information includes insurance products title, vehicle letter
Breath, premium information；
First computing module, for calculating the corresponding present valence of declaration form that can feed back to the client according to the policy information
It is worth threshold value；
Loading module, for the bonus information by the present value in bonus data library lower than the present of present value threshold value
Load is in the corresponding client of the client, and the bonus data library is used to store bonus information and bonus information is corresponding
Present value；
Second computing module for receiving at least one present of client's selection, and calculates respectively giving for client's selection
The aggregate value of the corresponding value of product；
Judgment module, for judging whether the aggregate value is less than or equal to the present value threshold value；
Determination module determines client's selection if being worth threshold value less than or equal to the present for the aggregate value
At least one present belongs to the present of the client.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811320261.6A CN109685543A (en) | 2018-11-07 | 2018-11-07 | Method, apparatus and computer equipment based on machine learning definite policy present |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811320261.6A CN109685543A (en) | 2018-11-07 | 2018-11-07 | Method, apparatus and computer equipment based on machine learning definite policy present |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109685543A true CN109685543A (en) | 2019-04-26 |
Family
ID=66185461
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811320261.6A Pending CN109685543A (en) | 2018-11-07 | 2018-11-07 | Method, apparatus and computer equipment based on machine learning definite policy present |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109685543A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110443575A (en) * | 2019-07-29 | 2019-11-12 | 苏州共创方程式企业管理有限公司 | A kind of customized conferring system and presentation method |
| CN110796557A (en) * | 2019-11-04 | 2020-02-14 | 泰康保险集团股份有限公司 | Data processing method and device, electronic equipment and computer readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20130130516A (en) * | 2012-05-22 | 2013-12-02 | (주)이패스 | Customized gift information providing system |
| CN106846041A (en) * | 2016-12-26 | 2017-06-13 | 携程计算机技术（上海）有限公司 | The distribution method and system of reward voucher |
| CN106875227A (en) * | 2017-02-25 | 2017-06-20 | 张元康 | A kind of Xian Xia shops rebate system based on IOS and Android |
| CN108256954A (en) * | 2017-10-31 | 2018-07-06 | 平安科技（深圳）有限公司 | A kind of choosing method and its terminal of Presents Giving scheme |
| CN108765174A (en) * | 2018-05-29 | 2018-11-06 | 泰康保险集团股份有限公司 | Insure rebating method and device |
-
2018
- 2018-11-07 CN CN201811320261.6A patent/CN109685543A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20130130516A (en) * | 2012-05-22 | 2013-12-02 | (주)이패스 | Customized gift information providing system |
| CN106846041A (en) * | 2016-12-26 | 2017-06-13 | 携程计算机技术（上海）有限公司 | The distribution method and system of reward voucher |
| CN106875227A (en) * | 2017-02-25 | 2017-06-20 | 张元康 | A kind of Xian Xia shops rebate system based on IOS and Android |
| CN108256954A (en) * | 2017-10-31 | 2018-07-06 | 平安科技（深圳）有限公司 | A kind of choosing method and its terminal of Presents Giving scheme |
| CN108765174A (en) * | 2018-05-29 | 2018-11-06 | 泰康保险集团股份有限公司 | Insure rebating method and device |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110443575A (en) * | 2019-07-29 | 2019-11-12 | 苏州共创方程式企业管理有限公司 | A kind of customized conferring system and presentation method |
| CN110796557A (en) * | 2019-11-04 | 2020-02-14 | 泰康保险集团股份有限公司 | Data processing method and device, electronic equipment and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8731977B1 (en) | System and method for analyzing and using vehicle historical data | |
| US20210350307A1 (en) | Intelligent Dealership Recommendation Engine | |
| CN109377286A (en) | The method, apparatus and computer equipment of insurance price are calculated based on prediction model | |
| US20150206249A1 (en) | Systems and methods for predicting the value of personal property | |
| US20140258044A1 (en) | Price scoring for vehicles | |
| US20200342478A1 (en) | Machine-Learning Data Processing System with Current Value Model | |
| KR102410777B1 (en) | An apparatus for improving way of product recommendation and a method thereof | |
| Baldwin et al. | Inflation, uncertainty, and investment | |
| CN109685543A (en) | Method, apparatus and computer equipment based on machine learning definite policy present | |
| KR20180117286A (en) | Method and Apparatus for pricing based on machine learning | |
| CN112559900A (en) | Product recommendation method and device, computer equipment and storage medium | |
| CN111406272A (en) | Configured price quotes with enhanced approval control | |
| CN113506143A (en) | Commodity discount generation method, device, equipment and computer readable storage medium | |
| US20040148241A1 (en) | Method of evaluating a portfolio of leased items | |
| CN109460964A (en) | Method, apparatus and computer equipment based on the more newly-generated vehicle insurance price list of data | |
| CN116503092A (en) | User reservation intention recognition method and device, electronic equipment and storage medium | |
| CN110490749A (en) | A kind of price fixing method and device | |
| CN112232945B (en) | Method and device for determining personal client credit | |
| CN114926261A (en) | Method and medium for predicting fraud probability of automobile financial user application | |
| US11126931B1 (en) | Intelligent dealership recommendation engine | |
| US20210027317A1 (en) | Inventory and structure finder | |
| CN113935673A (en) | Warehouse network optimization method and device based on inventory, computer equipment and storage medium | |
| CN110659754A (en) | Vehicle information determination method and device | |
| Liang et al. | Inventory control and replenishment with flexible delivery‐time upgrade | |
| KR102627498B1 (en) | Vehicle For Sale Verification Method Using Neural Network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |