CN107330719A - A kind of insurance products recommend method and system - Google Patents
A kind of insurance products recommend method and system Download PDFInfo
- Publication number
- CN107330719A CN107330719A CN201710433076.7A CN201710433076A CN107330719A CN 107330719 A CN107330719 A CN 107330719A CN 201710433076 A CN201710433076 A CN 201710433076A CN 107330719 A CN107330719 A CN 107330719A
- Authority
- CN
- China
- Prior art keywords
- label
- user
- insurance products
- product
- products
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
-
- 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)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Technology Law (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
Invention provides a kind of insurance products and recommends method and system, and methods described includes:The label of insurance products is extracted, is disassembled as different high and low layered each level labels, and is stored in tag database, the label is the characteristic attribute of insurance products;According to the data being stored with the tag database of each level label, label corresponding with the target product in new addition system is generated;According to the corresponding user behavior of different user to its recommended products label, so as to recommend the product matched indirectly.The present invention is by the way that user and tag match are got up, the insurance products combination that them can be recommended to be best suitable for user;Because effectively being matched between product, the effect of guarantee can be greatly reinforced;Insurance products are split out to the form of expression of inscape, system just can more precisely recommend the insurance products of closer to reality;It is known that different user groups, likes any label and products characteristics, it is thus possible to for the preference of different user, remove the product being more suitable for their recommendations.
Description
Technical field
The present invention relates to Products Show system regions, and in particular to a kind of insurance products recommend method and system.
Background technology
First, existing insurance products commending system is pushed to user, but protect per money according to insurance products self-characteristic completely
The inscape of dangerous product, such as insurance responsibility, insurance premium rate, premium delivery method, insurance period, insurance proceeds or insurance money
Mode of satisfaction all has greatly difference, it is difficult to only recommend client with excellent/bad two dimensions, because an insurance can include many
Feature, is to attract user's to be which feature so being not aware that.When user selects a certain product, technology backstage can not be very
Clearly recognize, which inscape user has an optimistic view of, or has an optimistic view of which factor combination of whole insurance products, therefore, user
Demand by backstage record analysis, can meet demand orientation type insurance market require.
Existing commending system can to recommend client should/possibility/product substantially liked, with such as demographics
Etc. statistical tool, but specifically accuracy rate is not high, and because insurance products are combined by each inscape, machinery is applied mechanically
The noise of combination can be very big, it is difficult to touches the core demand of client.
Second, personal insurance product on the market has thousands of moneys, and real fast-selling product only has over one hundred money, the result of recommendation
Substantially the product that this over one hundred money withstands market test is concentrated on, it is difficult to reach the individual cultivation effect in the face of thousand people thousand, lead
The use track data at family of applying can not react the preference of user well, more because product scope is limited to " quick-fried money ", recommend production
Product just can not jump this hundred products substantially, and lack of diversity causes new insurance products to appear on the market, and maximum probability can meet with cold open
Dynamic awkward condition.Wherein, cold start-up refers to that backstage does not have under the data precipitation situation of user, it is impossible to recommends or recommends to knot
It is really not accurate enough.Therefore, pay close attention to new product without client before for one, unless done specially treated, be otherwise difficult to be recommended.
The content of the invention
It is an object of the present invention in order to solve the above technical problems, provide.
In order to solve the above technical problems, the present invention is adopted the following technical scheme that:
The present invention provides a kind of insurance products and recommends method, and methods described includes:
S1, the label for extracting insurance products, are disassembled as different high and low layered each level labels, and be stored in label data
In storehouse, the label is the characteristic attribute of insurance products;
The data that S2, basis are stored with the tag database of each level label, generation and the target in new addition system
The corresponding label of product;
S3, according to the corresponding user behavior of different user to its recommended products label, so as to recommend what is matched indirectly
Product.
Further, insurance products as the aforementioned recommend method, described to disassemble insurance products for high and low layered label
Specially:The level label also includes optimization label;All inscapes of insurance products are carried out to disassemble each structure of acquisition
Into many forms of key element, outgoing label is extracted in every kind of form of expression;And successively by insurance products, optimization label, structure
Into key element and the form of expression with tree-shaped expansion, and identified with natural language;The optimization label is used for each inscape and table
Existing form carries out colloquial style performance.
Further, insurance products as the aforementioned recommend method, and the step S2 also includes:
Read the clause of target product;Participle is carried out to former clause sentence, keyword is extracted, then in the clause
Whether with the keyword corresponding label is had.
Further, insurance products as the aforementioned recommend method, in addition to:
Before the step S2, when system is initial, the label of typing originally each level of several insurance products, and handle
The corresponding former clause sentence of each label is stored in tag database and started for system jointly;
After the step S2, compared by tag database the label of each level disassembling out from target product and
Former clause sentence;Continue the former clause sentence content for adjusting label and matching, so as to improve the efficiency of machine evaluation and test.
Further, insurance products as the aforementioned recommend method, and the step S3 is specifically included:
When system obtains earlier user and accessed, user behavior is recorded, and user is arrived into user behavior storage
Behavior database;User is arrived into user's row to the behavior storage that the Products Show unit is made after displaying product result simultaneously
For in database;
The personal factor information of user is recorded, and the personal factor information is also stored into user behavior data storehouse;
After user behavior, the label of product and the personal factor information of user is combined, different users are obtained to mark
The preference of label;
When the user recorded in user and class of subscriber by system accesses again, according to the user to the inclined of label
It is good, the insurance products of the label of the user preference are found out, and according to weight of the user to different labels, recommend displaying relative
The product result answered.
The present invention also provides a kind of insurance products commending system, including:Product labelling unit, product analysis unit and product
Recommendation unit;The Product labelling unit disassembles unit including label;The product analysis system includes analysis module and label
Database;
The label disassembles the label that unit is used to extract insurance products, disassembles as different high and low layered each level marks
Label, and be stored in tag database, the label is the characteristic attribute of insurance products;
The analysis module is used for according to the data being stored with the tag database of each level label, generation and new addition
The corresponding label of target product in system;
Products Show unit is used for according to the corresponding user behavior of different user to its recommended products label, so as to push away indirectly
Recommend out the product matched.
Further, insurance products commending system as the aforementioned, it is described to disassemble insurance products for high and low layered label
Specially:The level label also includes optimization label;All inscapes of insurance products are carried out to disassemble each structure of acquisition
Into many forms of key element, outgoing label is extracted in every kind of form of expression;And successively by insurance products, optimization label, structure
Into key element and the form of expression with tree-shaped expansion, and identified with natural language;The optimization label is used for each inscape and table
Existing form carries out colloquial style performance.
Further, insurance products commending system as the aforementioned, the analysis module includes:
Clause sentence read module:Clause for reading target product;
Keyword extraction and comparison module:For carrying out participle to former clause sentence, keyword is extracted, it is then relatively more described
Whether with the keyword corresponding label is had in clause.
Further, insurance products commending system as the aforementioned, the product analysis unit also includes:Initialize mould
Block, contrast module and maintenance module;
The initialization module:For when system is initial, the label of typing originally each level of several insurance products, and
The corresponding former clause sentence of each label is stored in tag database jointly for system to start;
The contrast module:The label of each level disassembling out from target product for comparing by tag database and
Former clause sentence;
The maintenance module:For continuing the former clause sentence for adjusting label and matching on the basis of the contrast module
Content, so as to improve the efficiency of machine evaluation and test.
Further, insurance products commending system as the aforementioned, the Products Show unit includes:User behavior is recorded
Module, user profile logging modle, user preference acquisition module, Products Show module and user behavior data storehouse;
The user behavior logging modle:For when commending system obtains earlier user and accessed, recording user's row
For, and user behavior data storehouse is arrived into user behavior storage;
The user profile logging modle:Personal factor information for recording user, and by the personal factor information
Store user behavior data storehouse;The behavior that user is made to the Products Show unit after displaying product result simultaneously is deposited
Store up in user behavior data storehouse
The user preference acquisition module:For combining the personal dimension of user behavior, the label of insurance products and user
Spend after information, obtain preference of the different users to label;
The Products Show module:User for being recommended system record in user and class of subscriber is again introduced into
During commending system, according to preference of the user to label, the insurance products of the label of the user preference are found out, and according to user
To the weight of different labels, the product result for recommending displaying corresponding.
Compared with prior art, the present invention can be combined according to insurance label recommendations insurance:
User and tag match are got up, the insurance products combination that them can be recommended to be best suitable for user;Because product
Between effectively match, the effect of guarantee can be greatly reinforced;
Insurance products are split out to the form of expression of inscape, that is, will be splitted out the characteristics of entirely insurance, I
Platform and user terminal all therefrom income is huge, based on the commending system of this point, user's reality, personalized demand for insurance is just
It is embodied in the relation with each level label, system just can more precisely recommend the insurance products of closer to reality, suit the remedy to the case,
User terminal is also light, simply and intuitively bought required insurance products;
Analysis by backstage to user's representation data, can clearly know, which different user groups likes
A kind of label and products characteristics, it is thus possible to for the preference of different user, remove the product being more suitable for their recommendations;
Reversely customization:, can be according to the inclined of user by the present invention after enough user preference datas of insuring are collected into
It is good, and crowd characteristic, analyze the total demand of user;And according to obtained data, insurance is reversely customized to insurance company
Product.
The label for splitting into multi-layer will be insured, each insurance products averagely there are 50 labels, if having 1000 on the market
Individual multi-product, it is possible to obtain more than general 30000 independent insurance label;The insurance products data scale of magnanimity, quickly
Stream compression and various data type are the competitive advantage places of the present invention;Interacted by big data and specific insurer,
The people being truly realized matches with big data, meets the specific demand for insurance of insurer.
Brief description of the drawings
Fig. 1 is a kind of method flow schematic diagram of the invention;
A kind of specific implementation method schematic diagram that Fig. 2 is step S1 in Fig. 1;
A kind of specific implementation method schematic diagram that Fig. 3 is step S2 in Fig. 1;
A kind of specific implementation method schematic diagram that Fig. 4 is step S3 in Fig. 1;
Fig. 5 is insurance products-optimization label-tree-shaped expanded view of inscape-form of expression in the present invention;
Step S102 a kind of specific embodiment schematic diagram in Fig. 6;
Fig. 7 is that user in an embodiment of the present invention and system are shown with optimization label, inscape, interacting for the form of expression
It is intended to;
A kind of specific implementation method figure that Fig. 8 is step S3 in Fig. 1;
Fig. 9 is a kind of embodiment module map of present system.
The present invention is further illustrated with reference to the accompanying drawings and detailed description.
Embodiment
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is explicitly described, it is clear that described embodiment be the present invention
A part of embodiment, rather than whole embodiments.
The present invention provides a kind of insurance products and recommends method, and methods described includes:
S1, the label for extracting insurance products, are disassembled as different high and low layered each level tags, and be stored in label
In database 230, the label is the characteristic attribute of insurance products;
The data that S2, basis are stored with the tag database 230 of each level label, generation and the mesh in new addition system
Mark the corresponding label of product;
S3, according to the corresponding user behavior of different user to its recommended products label, so as to recommend what is matched indirectly
Product.
It is specially for high and low layered label by described disassemble insurance products in the Product labelling unit:By product
All inscapes carry out disassembling many forms that obtain each inscape, extract bid in every kind of form of expression
Label;And successively by product, inscape and the form of expression with tree-shaped expansion, and identified with natural language.
Fig. 2 is a kind of step S1 specific implementation method schematic diagram:
S101, insurance products are disassembled as high and low layered label, and is classified:
Described disassemble insurance products be specially for high and low layered label:All inscapes of product are disassembled
The many forms of each inscape are obtained, outgoing label is extracted in every kind of form of expression.Insurance products as described herein
To be present in the product in present system, target product is the new product added in present system, specifically, will insurance
Inscape (TAG A) broken of product;Each inscape has many forms of expression (TAG A-1,2,3), these
Possibility of the form of expression in different product carries out limit, extracts label characteristics code, the son as inscape (TAG A)
Collection, wherein the inscape of the insurance products mainly include insurance responsibility, insurance premium rate, premium delivery method, insurance period,
Insurance proceeds or insurance money mode of satisfaction etc., in addition to age at issue, the subdivision such as age, accessory risk, sex, hesitation phase of continuing insurance
Property customization factors.The form of expression may include:1st, insurance responsibility and the form of expression of exclusions;Such as the guarantee of serious illness insurance
Diseases range:Generally 25 kinds of major diseases of industry unified standard, general to exclude carcinoma in situ etc.;2nd, the performance shape of insurance premium rate
Formula:The insurance premium corresponding with particular risk can be simply interpreted as;3rd, the form of expression of premium delivery method:Typically can simply it divide
Hand over and buy wholesale by a definite date and hand over (disposably paying);4th, the form of expression of insurance period and observation period:Insurance contract life can be simply interpreted as
(contain continuation of insurance) during imitating date and insurer's obligation exemption;5th, the form of expression of insurance money (insurance proceeds) mode of satisfaction:Once
Get, get (such as annuity, season) by several times and the two is hybrid;6th, the manifestation mode of pay charge way:Once pay up and
Pay up, particularly may be divided into 3 years, 5 years, 10 years, 15 years, 20 years, 30 years by setting time.The corresponding number of tags of the every kind of form of expression
The how many speciality depending on insurance products of amount number, the characteristics of some products are covered be more, and corresponding label will be more.
It is preferred that, a product has 30~60 labels matched.It is used as system and the basic data of algorithm subsequent analysis.
S102. successively by product, inscape and the form of expression with tree-shaped expansion, and identified with natural language;The tree
Also include optimization label in shape expansion;And tree-shaped expansion mode is product, optimization label, inscape and the form of expression;Institute
Stating optimization label is used to carry out colloquial style performance to each inscape and the form of expression.Specifically, according to market, user group
Feature optimization inscape (TAG A) and the form of expression (TAG A-1,2,3), make it be rendered as user and can be used at natural language
The data of reason, as shown in figure 5, insurance products-optimization label-inscape (TAG A)-form of expression (TAG A-1,2,3) will
In tree-shaped expansion, the optimization label is the colloquial style expression to each inscape or the form of expression:Such as non deductible, cost performance
High, scope of insurance coverage is wide, year interest rate it is high, protect it is lifelong, enjoy dividend etc.;And shown with natural language, facilitate user understanding and
Select.
There is provided a kind of a kind of embodiment such as above-mentioned method S102 as shown in Figure 6:The insurance products are:Certain weight disease is protected
Danger, optimizing wherein two, label is:Non deductible and lifelong carefree;Optimization label correspondence inscape TAG includes:Deductible excess
And period insured;The optimization label corresponds to the form of expression:Non deductible and guarantor are lifelong;Therefore deleted and selected just according to the label
It is a kind of lifelong in full without the insurance products for acting certain the entitled weight disease insurance for paying for the amount of money.
S103. user is in commending system, and system can be by voice or text identification, identification user's request, and according to
The demand navigates to corresponding optimization label.
As shown in Figure 7:Optimize label, be the one layer label most with user mutual, be also that domestic consumer is recognized that
Level, this layer, which is mainly, collects demand tendency information of the different user to insurance products substantially.The label of the bottom is table
Existing form (TAG A-1,2,3), is adapted to the user known quite well to insurance, and this is also machine intelligence matching label used.
New insurance products appear on the market, during input coefficient is recognized, automatically extract out the form of expression (TAG A-1,2,3), are categorized into structure
Into in key element (TAG A) or even optimization label.
The benefit being so layered is:
1. the superiors optimize label, the research of user can be made with the development of market and product, and project team
Specific aim is changed, and is changed more frequent.Multilayer labels are divided, the design of multilayer labels is that bottom label has naturally various
Property and stability;The variation of a product changes the label that a product contains without multi-layer, it is only necessary to change upper label.
2. the selection according to user to label, system will set up user's portrait model, it is assessed for the ripe of insurance products
Degree is known, Products Show function is adjusted.
3. allowing user to get information about the internal relation of each label level, the demand for insurance of itself is preferably examined closely.
Fig. 3 is a kind of step S2 specific implementation method schematic diagram:
S201, when system is initial, the label of typing originally each level of several insurance products, and each label is relative
The former clause sentence answered is stored in tag database 230 and started for system jointly.
The data that S202, basis are stored with the tag database 230 of each level label, in generation and new addition system
The corresponding label of target product;Read the clause of target product;Participle is carried out to former clause sentence, keyword is extracted, then compares
Whether with the keyword corresponding label is had in the clause.Specifically:Number of the system in tag database 230
According to, in target product clause automatic identification its whether contain optimization label, inscape (TAG A), it is often more important that it is specific
The form of expression (TAG A-1,2,3) can be by accurate recording, and analyze and extract respective labels, classify.
A) reads the clause sentence of target product;
B) utilizes the theory and method of natural language processing (NLP), by former clause sentence participle, extracts keyword, then
Using TF-IDF, whether the common technology of the information retrieval such as cosine similarity algorithm and data mining is gone to compare among clause has this
The presence of label.The intuitively Rule of judgment of algorithm is:Provision content similarity is higher, there is the probability of corresponding label just
It is high.
S203, the label for comparing by tag database 230 each level for disassembling out from target product and former clause language
Sentence.Specifically:Backstage provides the function that random comparison, a full storehouse are compared, and is compared by tag database 230 from target product
Each level label, the clause sentence for disassembling out, adjust the sensitivity of system automatic identification, meet new Product labelling by chance, add
Into tag database 230.
Specifically, it is as described below to provide a specific embodiment for flow S201, S202, S203:
When a clause sentence:" within this contract life, insurant is from this execution of contract or multiple effect for Article 7
It is first after 180 days to occur and specified through our company or the medical institutions approved are made a definite diagnosis when suffering from major disease, our company
By the contained insured amount payment major disease insurance money of insurance policy, this validity of the contract is terminated at once.”
1. the clause sentence will be extracted lower column label:
(optimization label) observation period length-(TAG A) observation period -180 days (TAG A-n) observation periods;
Comprising a former clause sentence in tag database 230:Insurant is one from this execution of contract or multiple effect
It is first after 180 days to occur and specified through our company or the medical institutions approved are made a definite diagnosis when suffering from major disease ...
2. a pair former clause sentence is analyzed:
1) keyword is extracted after reading, participle:Insurant's/contract/comes into force/multiple effect/rise/mono- hundred eight ten day/...
2) technical finesse:
3. compare.
S204, continuation adjustment label and the former clause sentence content of matching on the basis of the contrast module, while on
State tag database 230:When tag extraction is more and more, tag system increasingly enrich, the evaluation and test effect of system and
Efficiency also more and more higher.
Fig. 4 is a kind of step S3 specific implementation method schematic diagram:
S301, when system obtains earlier user and accessed, record user behavior, and the user behavior storage arrived
User behavior data storehouse 350;User is arrived to the behavior storage that the Products Show unit is made after displaying product result and used
In family behavior database 350.
S302, the personal factor information for recording user, and user behavior data storehouse is arrived into personal factor information storage
350;
S303, after user behavior, the personal factor information of the label of product and user is combined, obtain different users
To the preference of label;
When S304, the user recorded in user and class of subscriber by system are again introduced into system, according to the user couple
The preference of label, finds out the product of the label of the user preference, and according to weight of the user to different labels, recommends displaying
Corresponding product result.
The step S301 further comprises:The user behavior storage is arrived behind user behavior data storehouse 350;Also by user
The behavior that the Products Show unit is made after displaying product result is stored into user behavior data storehouse 350.
It is illustrated in figure 8 a kind of step S3 specific embodiment flow chart:
A. when existing user comes and uses front end commending system, system can record their user behavior, and
It is stored to database.For example:The classification for the article that user clicks on, user see the preference of insurance evaluation and test, the letter that the intelligence seen is presented
Breath, the product that opened recommendation is combined, and the information searched for etc..
B. while can also record the personal factor information of user, including sex, age, identity, area and user's portrait etc..
C. in backstage, the behavioral data of user, the label of insurance and the personal information of user can be combined, with collaborative filtering
Algorithm (including based on user, based on content, or both all interactive algorithms, algorithm can increase according to different demands) go
Obtain preference of the different users to label.
D. when similar user (same user or same category of user) enters the commending system of the present invention, analysis
Module is extracted label from tag system first and performed an analysis, and then provides recommended products to commending system and product analysis unit respectively
Information and recommendation label information, then commending system just can be to the corresponding recommendation production of similar user displaying with product analysis unit
Product combine and recommend evaluation and test product tabulating result.
E. commending system can combine offline and online recommendation both of which, it is ensured that the promptness of recommendation.
F. the behavior that user makes for personalized recommendation, is deposited among user behavior data storehouse 350, root again
According to the feedback of active user, the effective percentage of commending system is able to know that, is then adjusted, logic is recommended in adjustment.
As shown in figure 9, the present invention also provides a kind of insurance products commending system, including:Product labelling unit 100, product
Analytic unit 200 and Products Show unit system 300;The Product labelling unit 100 disassembles unit 110 including label;The production
Product analysis system 200 includes analysis module 220 and tag database 230;
The label disassembles the label that unit 110 is used to extract insurance products, disassembles as different high and low layered each layers
Level tag, and be stored in tag database 230;
The analysis module 220 is used for according to the data that are stored with the tag database 230 of each level label, generation with
The corresponding label of target product in new addition system;
Products Show unit 300 is used for according to the corresponding user behavior of different user to its recommended products label, so that
Connect and recommend the product matched.
Further, insurance products commending system as the aforementioned, it is described to disassemble insurance products for high and low layered label
Specially:All inscapes of product are carried out to disassemble many forms for obtaining each inscape, in every kind of performance
Outgoing label is extracted in form;And successively by product, inscape and the form of expression with tree-shaped expansion, and identified with natural language.
Further, insurance products commending system as the aforementioned, also includes optimization label in the tree-shaped expansion;And
Tree-shaped expansion mode is product, optimization label, inscape and the form of expression;The optimization label be used for each inscape and
The form of expression carries out colloquial style performance.
Further, insurance products commending system as the aforementioned, the Product labelling unit 100 also includes:Voice and
Text identification module 120, corresponding optimization label is navigated to for recognizing user's request, and according to the demand.
Further, insurance products commending system as the aforementioned, the analysis module 220 includes:
Clause sentence read module 221:Clause for reading target product;
Keyword extraction and comparison module 222:For carrying out participle to former clause sentence, keyword is extracted, is then compared
Whether with the keyword corresponding label is had in the clause.
Further, insurance products commending system as the aforementioned, the product analysis unit 200 also includes:Initialization
Module 210, contrast module 240 and maintenance module 250;
The initialization module 210:For when system is initial, the mark of typing originally each level of several insurance products
Label, and the corresponding former clause sentence of each label is stored in tag database 230 for system startup jointly;
The contrast module 240:For comparing each level for disassembling out from target product by tag database 230
Label and former clause sentence;
The maintenance module 250:For continuing the former clause for adjusting label and matching on the basis of the contrast module
Sentence content, so as to improve the efficiency of machine evaluation and test.
Further, insurance products commending system as the aforementioned, the Products Show unit 300 includes:User behavior
Logging modle 310, user profile logging modle 320, user preference acquisition module 330, Products Show module 340 and user behavior
Database 350;
The user behavior logging modle 310:For when commending system obtains earlier user and accessed, recording user
Behavior, and user behavior data storehouse 350 is arrived into user behavior storage;
The user profile logging modle 320:Personal factor information for recording user, and the personal factor is believed
User behavior data storehouse 350 is arrived in breath storage;
The user preference acquisition module 330:For combining user behavior, the label of product and the personal factor of user
After information, preference of the different users to label is obtained;
The Products Show module 340:For being recommended the user of system record in user and class of subscriber again
During into commending system, according to preference of the user to label, the product of the label of the user preference is found out, and according to user
To the weight of different labels, the product result for recommending displaying corresponding.
Further, insurance products commending system as the aforementioned, the Products Show system includes offline and pushed away online
Recommend both of which;The user behavior logging modle 310 is additionally operable to:User is produced to the Products Show unit 300 in displaying
The behavior made after product result is stored into user behavior data storehouse 350.
The foregoing is only a specific embodiment of the invention, but protection scope of the present invention is not limited thereto, any
Those familiar with the art the invention discloses technical scope in, the change or replacement that can be readily occurred in, all should
It is included within the scope of the present invention.Therefore, protection scope of the present invention should it is described using scope of the claims as
It is accurate.
Claims (10)
1. a kind of insurance products recommend method, it is characterised in that methods described includes:
S1, the label for extracting insurance products, are disassembled as different high and low layered each level labels, and be stored in tag database
In, the label is the characteristic attribute of insurance products;
The data that S2, basis are stored with the tag database of each level label, generation and the target product in new addition system
Corresponding label;
S3, according to the corresponding user behavior of different user to its recommended products label, so as to recommend the product matched indirectly.
2. insurance products as claimed in claim 1 recommend method, it is characterised in that described to disassemble insurance products for high low layer
Secondary label is specially:The level label also includes optimization label;All inscapes of insurance products disassemble to obtain
The many forms of each inscape are obtained, outgoing label is extracted in every kind of form of expression;And successively by insurance products, optimization
Label, inscape and the form of expression are identified with tree-shaped expansion with natural language;The optimization label is used for will to each composition
Element and the form of expression carry out colloquial style performance.
3. insurance products as claimed in claim 1 recommend method, it is characterised in that the step S2 also includes:
Read the clause of target product;To former clause sentence carry out participle, extract keyword, then in the clause whether
There is label corresponding with the keyword.
4. insurance products as claimed in claim 1 recommend method, it is characterised in that also include:
Before the step S2, when system is initial, the label of typing originally each level of several insurance products, and each
The corresponding former clause sentence of label is stored in tag database and started for system jointly;
After the step S2, the label and former bar for each level for disassembling out from target product are compared by tag database
Money sentence;Continue the former clause sentence content for adjusting label and matching, so as to improve the efficiency of machine evaluation and test.
5. insurance products as claimed in claim 1 recommend method, it is characterised in that the step S3 is specifically included:
When system obtains earlier user and accessed, user behavior is recorded, and user behavior is arrived into user behavior storage
Database;
User is arrived into user behavior data to the behavior storage that the Products Show unit is made after displaying product result simultaneously
In storehouse;
The personal factor information of user is recorded, and the personal factor information is also stored into user behavior data storehouse;
After user behavior, the label of product and the personal factor information of user is combined, different users are obtained to label
Preference;
When the user recorded in user and class of subscriber by system accesses again, according to preference of the user to label, look for
Go out to have the insurance products of the label of the user preference, and according to weight of the user to different labels, recommend displaying corresponding
Product result.
6. a kind of insurance products commending system, it is characterised in that including:Product labelling unit, product analysis unit and product are pushed away
Recommend unit;The Product labelling unit disassembles unit including label;The product analysis system includes analysis module and number of tags
According to storehouse;The label, which disassembles unit, to be used to extract the labels of insurance products, is disassembled as different high and low layered each level labels,
And be stored in tag database, the label is the characteristic attribute of insurance products;
The analysis module is used for according to the data being stored with the tag database of each level label, and generation adds system with new
In the corresponding label of target product;
Products Show unit is used for according to the corresponding user behavior of different user to its recommended products label, so as to recommend indirectly
The product matched.
7. insurance products commending system as claimed in claim 6, it is characterised in that described to disassemble insurance products for high low layer
Secondary label is specially:The level label also includes optimization label;All inscapes of insurance products disassemble to obtain
The many forms of each inscape are obtained, outgoing label is extracted in every kind of form of expression;And successively by insurance products, optimization
Label, inscape and the form of expression are identified with tree-shaped expansion with natural language;The optimization label is used for will to each composition
Element and the form of expression carry out colloquial style performance.
8. insurance products commending system as claimed in claim 6, it is characterised in that the analysis module includes:
Clause sentence read module:Clause for reading target product;
Keyword extraction and comparison module:For carrying out participle to former clause sentence, keyword is extracted, then the clause
In whether have label corresponding with the keyword.
9. insurance products commending system as claimed in claim 6, it is characterised in that the product analysis unit also includes:Just
Beginningization module, contrast module and maintenance module;
The initialization module:For when system is initial, the label of typing originally each level of several insurance products, and each
The corresponding former clause sentence of individual label is stored in tag database and started for system jointly;
The contrast module:Label and former bar for comparing each level for disassembling out from target product by tag database
Money sentence;
The maintenance module:For continuing to adjust in the former clause sentence of label and matching on the basis of the contrast module
Hold, so as to improve the efficiency of machine evaluation and test.
10. insurance products commending system as claimed in claim 6, it is characterised in that the Products Show unit includes:User
Behavior record module, user profile logging modle, user preference acquisition module, Products Show module and user behavior data storehouse;
The user behavior logging modle:For when commending system obtains earlier user and accessed, recording user behavior, and
User behavior data storehouse is arrived into user behavior storage;
The user profile logging modle:Personal factor information for recording user, and the personal factor information is stored
To user behavior data storehouse;The behavior storage that user is made to the Products Show unit after displaying product result simultaneously is arrived
In user behavior data storehouse
The user preference acquisition module:For combining user behavior, the label of insurance products and the personal factor of user letter
After breath, preference of the different users to label is obtained;
The Products Show module:User for being recommended system record in user and class of subscriber is again introduced into recommendation
During system, according to preference of the user to label, the insurance products of the label of the user preference are found out, and according to user to not
The weight of same label, the product result for recommending displaying corresponding.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710433076.7A CN107330719A (en) | 2017-06-09 | 2017-06-09 | A kind of insurance products recommend method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710433076.7A CN107330719A (en) | 2017-06-09 | 2017-06-09 | A kind of insurance products recommend method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107330719A true CN107330719A (en) | 2017-11-07 |
Family
ID=60194734
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710433076.7A Pending CN107330719A (en) | 2017-06-09 | 2017-06-09 | A kind of insurance products recommend method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107330719A (en) |
Cited By (27)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108053324A (en) * | 2017-11-28 | 2018-05-18 | 平安养老保险股份有限公司 | Adjuster method, apparatus, computer equipment and storage medium |
CN108133012A (en) * | 2017-12-22 | 2018-06-08 | 新奥(中国)燃气投资有限公司 | A kind of label setting method and device |
CN108446990A (en) * | 2018-03-15 | 2018-08-24 | 全民福网络科技有限公司 | Insurance service is in line platform |
CN108460654A (en) * | 2018-02-06 | 2018-08-28 | 闽南师范大学 | A kind of insurance products recommendation method, system, medium and equipment based on binary tree |
CN108510402A (en) * | 2018-06-06 | 2018-09-07 | 中国平安人寿保险股份有限公司 | Insurance kind information recommendation method, device, computer equipment and storage medium |
CN108665316A (en) * | 2018-05-14 | 2018-10-16 | 中国平安人寿保险股份有限公司 | Insurance products method for pushing, device, equipment and computer readable storage medium |
CN108711110A (en) * | 2018-08-14 | 2018-10-26 | 中国平安人寿保险股份有限公司 | Insurance products recommend method, apparatus, computer equipment and storage medium |
CN108769143A (en) * | 2018-05-11 | 2018-11-06 | 平安科技(深圳)有限公司 | Information push method and device, storage medium, electronic equipment |
CN108961087A (en) * | 2018-07-13 | 2018-12-07 | 众安在线财产保险股份有限公司 | Insure recommended method, device, computer equipment and computer readable storage medium |
CN108959618A (en) * | 2018-07-18 | 2018-12-07 | 北京欣欣苹果网络科技有限公司 | Internet information Collecting and dealing method and apparatus |
CN109146610A (en) * | 2018-07-16 | 2019-01-04 | 众安在线财产保险股份有限公司 | It is a kind of intelligently to insure recommended method, device and intelligence insurance robot device |
CN109190925A (en) * | 2018-08-13 | 2019-01-11 | 中国平安人寿保险股份有限公司 | Declaration form recommended method, device, computer equipment and storage medium |
CN109285078A (en) * | 2018-08-20 | 2019-01-29 | 中国平安人寿保险股份有限公司 | Give the method, apparatus, computer equipment and storage medium of insurance |
CN109711997A (en) * | 2018-08-20 | 2019-05-03 | 平安普惠企业管理有限公司 | Insurance business method for pushing, device, equipment and readable storage medium storing program for executing |
CN109727085A (en) * | 2018-06-01 | 2019-05-07 | 平安普惠企业管理有限公司 | Insurance sale method, apparatus, equipment and readable storage medium storing program for executing |
CN109727139A (en) * | 2018-01-12 | 2019-05-07 | 中国平安财产保险股份有限公司 | Insure set meal customization method, device, equipment and readable storage medium storing program for executing |
CN109961370A (en) * | 2019-01-29 | 2019-07-02 | 宜信博诚保险销售服务(北京)股份有限公司 | A kind of the interests methods of exhibiting and device of insurance products |
CN110033381A (en) * | 2019-01-11 | 2019-07-19 | 阿里巴巴集团控股有限公司 | A kind of insurance recommended method and device for group visually impaired |
CN110111183A (en) * | 2019-05-06 | 2019-08-09 | 北京车薄荷科技有限公司 | A kind of Products Show method and device |
CN110111173A (en) * | 2019-04-12 | 2019-08-09 | 中国平安人寿保险股份有限公司 | Insurance products push control method, device, computer equipment and storage medium |
CN110288484A (en) * | 2019-04-02 | 2019-09-27 | 上海瀚之友信息技术服务有限公司 | A kind of the user's recommended method and system of the insurance classification based on big data platform |
CN110689446A (en) * | 2019-10-30 | 2020-01-14 | 海腾保险代理有限公司 | Combined insurance recommendation method and device |
CN110969521A (en) * | 2018-09-30 | 2020-04-07 | 重庆小雨点小额贷款有限公司 | Loan method, loan device, loan server and loan storage medium |
WO2020164332A1 (en) * | 2019-02-12 | 2020-08-20 | 阿里巴巴集团控股有限公司 | Insurance service processing method, apparatus and device |
CN112464660A (en) * | 2020-11-25 | 2021-03-09 | 平安医疗健康管理股份有限公司 | Text classification model construction method and text data processing method |
CN112463931A (en) * | 2020-12-11 | 2021-03-09 | 中国人寿保险股份有限公司 | Intelligent analysis method for insurance product clauses and related equipment |
CN112487287A (en) * | 2020-11-26 | 2021-03-12 | 深圳韦格纳医学检验实验室 | Method for recommending severe illness risk by using gene detection result and questionnaire |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103473354A (en) * | 2013-09-25 | 2013-12-25 | 焦点科技股份有限公司 | Insurance recommendation system framework and insurance recommendation method based on e-commerce platform |
CN104463630A (en) * | 2014-12-11 | 2015-03-25 | 新一站保险代理有限公司 | Product recommendation method and system based on characteristics of online shopping insurance products |
US9280252B1 (en) * | 2013-03-08 | 2016-03-08 | Allstate Insurance Company | Configuring an application task list of an application based on previous selections of application tasks |
CN106204202A (en) * | 2016-06-29 | 2016-12-07 | 百度在线网络技术(北京)有限公司 | A kind of vehicle insurance information recommendation method and device |
CN106503025A (en) * | 2015-09-08 | 2017-03-15 | 北京搜狗科技发展有限公司 | Method and system is recommended in a kind of application |
CN106649316A (en) * | 2015-10-29 | 2017-05-10 | 北京国双科技有限公司 | Video pushing method and device |
-
2017
- 2017-06-09 CN CN201710433076.7A patent/CN107330719A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9280252B1 (en) * | 2013-03-08 | 2016-03-08 | Allstate Insurance Company | Configuring an application task list of an application based on previous selections of application tasks |
CN103473354A (en) * | 2013-09-25 | 2013-12-25 | 焦点科技股份有限公司 | Insurance recommendation system framework and insurance recommendation method based on e-commerce platform |
CN104463630A (en) * | 2014-12-11 | 2015-03-25 | 新一站保险代理有限公司 | Product recommendation method and system based on characteristics of online shopping insurance products |
CN106503025A (en) * | 2015-09-08 | 2017-03-15 | 北京搜狗科技发展有限公司 | Method and system is recommended in a kind of application |
CN106649316A (en) * | 2015-10-29 | 2017-05-10 | 北京国双科技有限公司 | Video pushing method and device |
CN106204202A (en) * | 2016-06-29 | 2016-12-07 | 百度在线网络技术(北京)有限公司 | A kind of vehicle insurance information recommendation method and device |
Cited By (39)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108053324A (en) * | 2017-11-28 | 2018-05-18 | 平安养老保险股份有限公司 | Adjuster method, apparatus, computer equipment and storage medium |
CN108133012A (en) * | 2017-12-22 | 2018-06-08 | 新奥(中国)燃气投资有限公司 | A kind of label setting method and device |
CN108133012B (en) * | 2017-12-22 | 2022-01-18 | 新奥(中国)燃气投资有限公司 | Label setting method and device |
CN109727139A (en) * | 2018-01-12 | 2019-05-07 | 中国平安财产保险股份有限公司 | Insure set meal customization method, device, equipment and readable storage medium storing program for executing |
CN108460654A (en) * | 2018-02-06 | 2018-08-28 | 闽南师范大学 | A kind of insurance products recommendation method, system, medium and equipment based on binary tree |
CN108446990A (en) * | 2018-03-15 | 2018-08-24 | 全民福网络科技有限公司 | Insurance service is in line platform |
CN108769143B (en) * | 2018-05-11 | 2021-07-30 | 平安科技(深圳)有限公司 | Message pushing method and device, storage medium and electronic equipment |
CN108769143A (en) * | 2018-05-11 | 2018-11-06 | 平安科技(深圳)有限公司 | Information push method and device, storage medium, electronic equipment |
CN108665316A (en) * | 2018-05-14 | 2018-10-16 | 中国平安人寿保险股份有限公司 | Insurance products method for pushing, device, equipment and computer readable storage medium |
CN108665316B (en) * | 2018-05-14 | 2023-06-30 | 中国平安人寿保险股份有限公司 | Insurance product pushing method, apparatus, equipment and computer readable storage medium |
CN109727085A (en) * | 2018-06-01 | 2019-05-07 | 平安普惠企业管理有限公司 | Insurance sale method, apparatus, equipment and readable storage medium storing program for executing |
CN108510402A (en) * | 2018-06-06 | 2018-09-07 | 中国平安人寿保险股份有限公司 | Insurance kind information recommendation method, device, computer equipment and storage medium |
CN108961087A (en) * | 2018-07-13 | 2018-12-07 | 众安在线财产保险股份有限公司 | Insure recommended method, device, computer equipment and computer readable storage medium |
CN109146610A (en) * | 2018-07-16 | 2019-01-04 | 众安在线财产保险股份有限公司 | It is a kind of intelligently to insure recommended method, device and intelligence insurance robot device |
CN108959618A (en) * | 2018-07-18 | 2018-12-07 | 北京欣欣苹果网络科技有限公司 | Internet information Collecting and dealing method and apparatus |
CN108959618B (en) * | 2018-07-18 | 2021-12-24 | 北京欣欣苹果网络科技有限公司 | Internet information collecting and processing method and device |
CN109190925B (en) * | 2018-08-13 | 2023-09-29 | 中国平安人寿保险股份有限公司 | Policy recommendation method, device, computer equipment and storage medium |
CN109190925A (en) * | 2018-08-13 | 2019-01-11 | 中国平安人寿保险股份有限公司 | Declaration form recommended method, device, computer equipment and storage medium |
CN108711110A (en) * | 2018-08-14 | 2018-10-26 | 中国平安人寿保险股份有限公司 | Insurance products recommend method, apparatus, computer equipment and storage medium |
CN108711110B (en) * | 2018-08-14 | 2023-06-23 | 中国平安人寿保险股份有限公司 | Insurance product recommendation method, apparatus, computer device and storage medium |
CN109285078A (en) * | 2018-08-20 | 2019-01-29 | 中国平安人寿保险股份有限公司 | Give the method, apparatus, computer equipment and storage medium of insurance |
CN109711997A (en) * | 2018-08-20 | 2019-05-03 | 平安普惠企业管理有限公司 | Insurance business method for pushing, device, equipment and readable storage medium storing program for executing |
CN110969521A (en) * | 2018-09-30 | 2020-04-07 | 重庆小雨点小额贷款有限公司 | Loan method, loan device, loan server and loan storage medium |
CN110969521B (en) * | 2018-09-30 | 2023-09-12 | 重庆小雨点小额贷款有限公司 | Loan method, loan device, server and storage medium |
CN110033381A (en) * | 2019-01-11 | 2019-07-19 | 阿里巴巴集团控股有限公司 | A kind of insurance recommended method and device for group visually impaired |
WO2020143411A1 (en) * | 2019-01-11 | 2020-07-16 | 阿里巴巴集团控股有限公司 | Insurance recommendation method and device for visual impaired group |
CN109961370A (en) * | 2019-01-29 | 2019-07-02 | 宜信博诚保险销售服务(北京)股份有限公司 | A kind of the interests methods of exhibiting and device of insurance products |
WO2020164332A1 (en) * | 2019-02-12 | 2020-08-20 | 阿里巴巴集团控股有限公司 | Insurance service processing method, apparatus and device |
CN110288484A (en) * | 2019-04-02 | 2019-09-27 | 上海瀚之友信息技术服务有限公司 | A kind of the user's recommended method and system of the insurance classification based on big data platform |
CN110288484B (en) * | 2019-04-02 | 2022-12-13 | 上海瀚之友信息技术服务有限公司 | Insurance classification user recommendation method and system based on big data platform |
CN110111173A (en) * | 2019-04-12 | 2019-08-09 | 中国平安人寿保险股份有限公司 | Insurance products push control method, device, computer equipment and storage medium |
CN110111183A (en) * | 2019-05-06 | 2019-08-09 | 北京车薄荷科技有限公司 | A kind of Products Show method and device |
CN110689446A (en) * | 2019-10-30 | 2020-01-14 | 海腾保险代理有限公司 | Combined insurance recommendation method and device |
CN112464660B (en) * | 2020-11-25 | 2023-02-07 | 深圳平安医疗健康科技服务有限公司 | Text classification model construction method and text data processing method |
CN112464660A (en) * | 2020-11-25 | 2021-03-09 | 平安医疗健康管理股份有限公司 | Text classification model construction method and text data processing method |
CN112487287A (en) * | 2020-11-26 | 2021-03-12 | 深圳韦格纳医学检验实验室 | Method for recommending severe illness risk by using gene detection result and questionnaire |
CN112487287B (en) * | 2020-11-26 | 2024-03-22 | 深圳韦格纳医学检验实验室 | Method for recommending serious diseases by using gene detection result and questionnaire |
CN112463931A (en) * | 2020-12-11 | 2021-03-09 | 中国人寿保险股份有限公司 | Intelligent analysis method for insurance product clauses and related equipment |
CN112463931B (en) * | 2020-12-11 | 2024-05-28 | 中国人寿保险股份有限公司 | Intelligent analysis method and related equipment for insurance product clauses |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107330719A (en) | A kind of insurance products recommend method and system | |
US11095594B2 (en) | Location resolution of social media posts | |
US11405344B2 (en) | Social media influence of geographic locations | |
AU2011242753B2 (en) | Product synthesis from multiple sources | |
US20150278355A1 (en) | Temporal context aware query entity intent | |
Benchimol et al. | Text mining methodologies with R: An application to central bank texts | |
US20090265328A1 (en) | Predicting newsworthy queries using combined online and offline models | |
WO2018040069A1 (en) | Information recommendation system and method | |
US9344507B2 (en) | Method of processing web access information and server implementing same | |
CN105378730A (en) | Social media content analysis and output | |
JP6056610B2 (en) | Text information processing apparatus, text information processing method, and text information processing program | |
Belem et al. | “Fixing the curse of the bad product descriptions”–Search-boosted tag recommendation for E-commerce products | |
CN110598126B (en) | Cross-social network user identity recognition method based on behavior habits | |
Ashraf et al. | Personalized news recommendation based on multi-agent framework using social media preferences | |
CN113407729B (en) | Judicial-oriented personalized case recommendation method and system | |
CN113591476A (en) | Data label recommendation method based on machine learning | |
CN110209944A (en) | A kind of stock analysis teacher recommended method, device, computer equipment and storage medium | |
CN103514237B (en) | A kind of method and system obtaining user and Document personalization feature | |
JP6260678B2 (en) | Information processing apparatus, information processing method, and information processing program | |
US20220222715A1 (en) | System and method for detecting and analyzing discussion points from written reviews | |
Zhu et al. | Identifying and modeling the dynamic evolution of niche preferences | |
Krishnamoorthy et al. | TV shows popularity and performance prediction using CNN algorithm | |
US20180033056A1 (en) | Competitor trend-based social content ideation | |
CN112307298A (en) | Method and device for generating personal brand label | |
Xia et al. | An application of recommender system with mingle-topn algorithm on b2b platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171107 |
|
RJ01 | Rejection of invention patent application after publication |