CN106600369A - Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification - Google Patents
Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification Download PDFInfo
- Publication number
- CN106600369A CN106600369A CN201611129818.9A CN201611129818A CN106600369A CN 106600369 A CN106600369 A CN 106600369A CN 201611129818 A CN201611129818 A CN 201611129818A CN 106600369 A CN106600369 A CN 106600369A
- Authority
- CN
- China
- Prior art keywords
- data
- feature
- client
- financial product
- module
- 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
-
- 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
- G06Q40/02—Banking, e.g. interest calculation or account maintenance
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computational Mathematics (AREA)
- Artificial Intelligence (AREA)
- Algebra (AREA)
- Probability & Statistics with Applications (AREA)
- Technology Law (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention discloses a real-time recommendation system and method of financial products of banks based on Naive Bayesian classification wherein the method comprises: acquiring the data of a bank business system, a peripheral system, and an internet/mobile internet; processing and storing the data in a basic database; conducting characteristic processing to the data attributes of the customers in a certain area; converting for data characteristics to develop a characteristic database; classifying the data in the characteristic database as a training data set and a verification data set which are used respectively for the training and estimating of a Naive Bayesian classification algorithm model; obtaining a model base after several times of iteration; identifying the characteristics of customers and in combination with the Naive Bayesian classification algorithm model base, matching the customers with the financial products they are mostly likely to purchase in real time; and recommending to the customers the financial products that are forecasted and matched by the models. The recommendation system and method of the invention are capable of helping customers to find out suitable products, therefore, increasing the user experience and the bonding effect of the customers to the financial products, lifting up the transferring rate of the products and making the bank more competitive.
Description
Technical field
The present invention relates to a kind of bank finance product real-time recommendation system and method, more particularly to it is a kind of based on simple pattra leaves
The bank finance product real-time recommendation system and method for this classification.
Background technology
The development of the technologies such as the Internet, big data, artificial intelligence, changes the competition situation of banking；Pursue ultimate attainment
Consumer's Experience, emphasize that data-driven operation becomes new feature.At present, the financial product marketing of most of banks is built upon passing
On system manipulation type marketing system, general operation flow is first to carry out statistical analysiss to data such as client, product, transaction, then
A collection of client is filtered out with reference to business rule, then by phone or note promoting certain product that bank specifies.This battalion
Pin pattern in ageing and automaticity can not rapidly adapt to turn of the market and carry out real time individual clothes to client
The requirement of business, which has the disadvantage that：
1st, full client's marketing can not be supported.The customer data of current bank full dose is very huge, up to TB levels, traditional marketing system
Technical Architecture is in memory capacity, the storage and processing of client's full dose data cannot be met in processing characteristics, in time window
On the requirement that cannot also meet business to data mart modeling and calculate in real time, therefore, traditional marketing system can only be supported to part
The marketing of client.
2nd, full product marketing can not be supported.The financial product of bank is very more, has fund, financing, noble metal, insurance etc. each
Class product, each class include several products again；In the face of up to thousand of kinds of product, client cannot quickly select a kind of being adapted to certainly
, there is " product information overload " in oneself product；And in terms of bank, due to being short of hands or business is currently needed for, Zhi Nengxuan
Take certain several product to be marketed, and most products are not promoted to the chance of client, there is " long-tail product ".
3rd, full channel marketing can not be supported.Traditional Marketing system, usually marketing personnel are produced by phone or note
Product are marketed, and as telephone fraud, short message fraud are multiple, bank client compares conflict to phone, note, by phone, note etc.
The efficiency marketed by channel is very low, therefore Traditional Marketing system does not give full play to full channel and is particularly mobile Internet canal
Effect of the road in marketing.
4th, real time individual precision marketing can not be supported.Traditional marketing system, simply by traditional analytic statisticss side
Method carries out product batch with reference to business rule screening client and markets, and the feature, hobby, trading activity not according to client is carried out
The in-depth analysis and excavation of line, it is impossible to real time individual precision marketing is carried out to client and service is customized.
To sum up, it is necessary to design a kind of bank finance product real-time recommendation system and method to make up drawbacks described above.
The content of the invention
The present invention proposes a kind of bank finance product real-time recommendation system and method based on Naive Bayes Classification, its solution
" product information overload ", marketing response delay and " long-tail product that Jue Liao banks Traditional Marketing system and Traditional Marketing pattern are present
The defect of product ".The present invention adopts distributed big data platform, distributed memory computing technique and Naive Bayes Classification Algorithm structure
Build and form, client can be helped to be quickly found out suitable product, so as to improving customer experience, increased the viscosity of client, carry
Rise conversion rate of products, improve the competitiveness of bank.
The technical scheme is that what is be achieved in that：
The present invention discloses a kind of real-time recommendation method of the bank finance product real-time recommendation system based on Naive Bayes Classification,
Which comprises the steps：（S01）Collection banking system, peripheral system, the data of the Internet/mobile Internet, and process
With storage to basic database；（S02）Characteristic processing is carried out to the data attribute of full dose client, data characteristicses is converted to and is formed spy
Levy storehouse；（S03）Data in feature database are divided into into training dataset and checking data set, Naive Bayes Classification calculation is respectively used to
The training and assessment of method model, Jing successive ignitions obtain model library；（S04）Identification client characteristics, and with reference to naive Bayesian point
The financial product that class algorithm model storehouse real-time matching client most possibly buys；（S05）The gold that model prediction module is matched
Melt client of the product real time propelling movement to application end.
Wherein, step（S01）Comprise the steps：（S11）Visitor is gathered by offline acquisition mode and online acquisition mode
Family attribute data, financial product attribute data and customer action data；（S12）To step（S11）In the data that collect carry out
Quality analysiss, data scrubbing and Data Integration.
Wherein, step（S11）The Data Source of middle collection includes：1）The customer information of kernel business of bank system, deposit loan
Money, the credit card, fund, financing and insurance data；2）The wage of local characteristic service system, charges for water and electricity, management for infrastructure fee and logical
Letter takes transaction data；3）People's row collage-credit data of peripheral system；4）The Web bank of the Internet/mobile Internet, wechat bank,
Mobile banking APP, social media and electric quotient data.
Wherein, step（S02）Comprise the steps：（S21）All client properties are analyzed, suitable data are selected
Attribute derives new feature as candidate feature while calculating to candidate feature or various data splittings；（S22）
According to characteristic and the requirement of Naive Bayes Classification Algorithm model, to step（S11）In candidate feature carry out characteristic processing；
（S23）Select to obtain the feature set of model and algorithm top performance；（S24）Reduce eigenmatrix dimension, and form feature database carrying
Supply model training module is used.
Wherein, step（S22）The engineering method of middle characteristic processing includes following process step：（S221）By discretization side
Method carries out discrete processes to multiple continuouss features；（S222）Continuouss feature is carried out at nondimensionalization by method for normalizing
Reason；（S223）Qualitative features are converted to by quantitative characteristic by mute coded method.
Wherein, step（S222）In method for normalizing include min-max methods and z-score methods.
Wherein, step（S03）Comprise the steps：（S31）It is random to extract 80% data as training set from feature database,
Use for training module；In feature database, remaining 20% data are used for model evaluation module as checking collection；（S32）It is defeated
Enter training set, the condition for selling probability each financial product corresponding with every kind of feature for calculating and preserving each financial product is general
Rate, and the financial product recommended models storehouse based on Naive Bayes Classification is set up for incorporating parametric；（S33）Input validation collection,
Client's purchase financial product is predicted, concentrates the financial product of actual purchase to calculate hit according to predicting the outcome and verifying
Rate, counts evaluation metricses and carrys out assessment models.
Wherein, step（S04）Comprise the steps：（S41）To send using AM access module customer's identity card number,
Bank's card number, fingerprint image, face-image carry out client's identification, obtain client characteristics collection；（S42）Existed according to the feature set of client
The conditional probability of each financial product of every kind of feature correspondence is read in model library, in conjunction with each financial product sell probability and
Parameter calculates financial product classification and ordination value；Corresponding financial product collection is predicted according to financial product classification and ordination value；（S43）
Prediction financial product collection out is screened, final financial product is obtained, and by being pushed to visitor using AM access module
Family.
Wherein, step（S05）Client of the middle financial product real time propelling movement that model prediction module is matched to application end
When, the push mode for using includes Web Service, HTML5, JavaScript, REST and Socket, and which supports site row
The financial product of team's machine, site self-service equipment, wifi, Web bank, Mobile banking APP and handset Wechat Bank application is recommended to connect
Enter function.
Invention additionally discloses a kind of bank finance product real-time recommendation system based on Naive Bayes Classification, which includes number
According to acquisition module, Feature Engineering module, model training module, model prediction module, using AM access module and data repository；Number
Include basic database, feature database and model library according to thesauruss；Data acquisition module is used to gather banking system, periphery system
System, the data of the Internet/mobile Internet, and process and store basic database；Feature Engineering module is for full dose visitor
The data attribute at family carries out characteristic processing, is converted to data characteristicses and forms feature database；Model training module is for by feature database
Data be divided into training dataset and checking data set, be respectively used to the training and assessment of Naive Bayes Classification Algorithm model,
Jing successive ignitions obtain model library；Model prediction module is used to recognize client characteristics, and combines Naive Bayes Classification Algorithm mould
The financial product that type storehouse real-time matching client most possibly buys；Using AM access module be used for obtain client identification card number or
Bank's card number simultaneously submits to model prediction module, then the financial product real time propelling movement that model prediction module is matched is to application end
Client.
Compared with prior art, the invention has the advantages that：
The present invention realizes the smooth extension of system expensive storage and the linear lifting of systematic function using distributed big data platform,
Solve, meet full dose client storage and extension demand comprehensively.
The present invention adopts big data distributed computing framework and distributed memory computing technique greatly to improve offline number
According to processing and the online performance for calculating in real time, satisfaction completes the processing of full dose customer data and calculating in real time in reasonable time window
Requirement.
The present invention replaces artificial operation by Machine automated marketing flow process, replaces artificial judgment by algorithm engine, entirely
Weather uninterrupted operation, full Products Show, full channel services client.
The present invention replaces artificial judgment, system to use Naive Bayes Classification Algorithm by algorithm engine, special according to client
Levying the principle that similarity is higher, concern financial product is more close carries out product screening, realizes the personalization essence to different clients
Quasi- marketing.
Description of the drawings
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
Accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is frame diagram of the present invention based on the bank finance product real-time recommendation system of Naive Bayes Classification.
Fig. 2 is the schematic diagram of data acquisition of the present invention.
Fig. 3 is the schematic diagram of feature of present invention engineering.
Fig. 4 is the schematic diagram of model training of the present invention.
Fig. 5 is the schematic diagram of model prediction of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than the embodiment of whole.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
In order to contribute to and clarify the description of subsequent embodiment, carry out specifically in the specific embodiment to the present invention
Before bright, part term is explained, following explanation is applied to this specification and claims.
The hadoop occurred in the present invention is a distributed system architecture developed by Apache funds club, its
The Chinese meaning is big data platform；Other English words are code, do not represent other in all senses.
Referring to figs. 1 to Fig. 5, the present invention discloses a kind of bank finance product real-time recommendation system based on Naive Bayes Classification
The real-time recommendation method of system, which comprises the steps：（S01）Collection banking system, peripheral system, the Internet/movement are mutually
The data of networking, and process and store basic database；（S02）Characteristic processing is carried out to the data attribute of full dose client, is turned
It is changed to data characteristicses and forms feature database；（S03）Data in feature database are divided into into training dataset and checking data set, are used respectively
In the training and assessment of Naive Bayes Classification Algorithm model, Jing successive ignitions obtain model library；（S04）Identification client characteristics,
And the financial product most possibly bought with reference to Naive Bayes Classification Algorithm model library real-time matching client；（S05）By model
Client of the financial product real time propelling movement that prediction module is matched to application end.
Wherein, step（S01）Comprise the steps：（S11）Visitor is gathered by offline acquisition mode and online acquisition mode
Family attribute data, financial product attribute data and customer action data；（S12）To step（S11）In the data that collect carry out
Quality analysiss, data scrubbing and Data Integration.Step（S11）In offline acquisition mode refer to origin system generate data file,
Then pass through file transmitting software（Such as FTP）It is sent to the system；Online acquisition mode refers to channel and online and real time data acquisition to this
System, such as captures software by delta data（Such as CDC）The incremental data of collection relevant database, is located in real time using flow data
Reason software（Such as Kafka+Spark Streaming）The data such as collection daily record, website purchase click steam.For bank core industry
Business system, local characteristic service system, using offline acquisition mode；For peripheral system and the number of the Internet/mobile Internet
According to gathering two ways as the case may be using online acquisition or offline.Step（S12）In, after data acquisition, enter line number
According to a series of data prediction such as quality analysiss, data scrubbing and Data Integration；Correctness first to data, accuracy, one
Cause property, integrity, whether meet the aspects such as business rule and checked；Then to existing, numerical exception, type are abnormal, run counter to industry
The data of the problems such as business rule, Data duplication carry out depth cleaning；Pass through client association information again（Such as identity card, phone number
Deng）Customer data from each system is integrated, normalized full dose customer basis data base is formed.
Wherein, step（S11）The Data Source of middle collection includes：1）The customer information of kernel business of bank system, deposit loan
Money, the credit card, fund, financing and insurance data；2）The wage of local characteristic service system, charges for water and electricity, management for infrastructure fee and logical
Letter takes transaction data；3）People's row collage-credit data of peripheral system；4）The Web bank of the Internet/mobile Internet, wechat bank,
Mobile banking APP, social media and electric quotient data.
Step of the present invention（S02）The items of Feature Engineering are carried out using Feature Engineering related software and method and with reference to coding
Process, be the process that data attribute is converted to the data characteristicses that fit algorithm model is used, from extracting data and algorithm mould
The validity feature of type height correlation, removes the data attribute of unrelated, redundancy, reduces noise jamming, is that the brief, computing of structure is fast
Spend the only stage which must be passed by of the Naive Bayes Classification Algorithm model of fast, readily understood, easy care；Step（S02）Quality directly concern
The effect of model prediction, is a very important ring.Step（S02）Comprise the steps：（S21）According to domain knowledge and Jing
Test, all client properties are analyzed by brainstorming, suitable data attribute is selected as candidate feature.Simultaneously
Candidate feature or various data splittings are carried out calculating with statistical rules or machine learning techniques and derive new feature, such as
Sex statistics is carried out to fund purchaser record and obtains fund sex characteristicss；（S22）According to Naive Bayes Classification Algorithm model
Characteristic and requirement, to step（S11）In candidate feature carry out characteristic processing；（S23）Select to obtain model and the best property of algorithm
The feature set of energy；（S24）Reduce eigenmatrix dimension, and form feature database being supplied to model training module to use.
Step（S21）In candidate feature include：Customer ID, type of credential, passport NO., name, sex, working year
Limit, home address, CompanyAddress, home address postcode, mobile phone, Email, date of birth, age, marital status,
Go through, work unit, occupation, site of opening an account, the date of opening an account, the time limit of opening an account, whether bank employee, the balance of deposits, loan balance, loan
Money total value, the accrediting amount, personal total assets, personal total liability, other client's acceptable level of risk, fund generic attribute, financing generic
Property, precious metal attribute, the insurance 80 multinomial data attributes such as generic attribute, constitute customer information width table.
Step（S23）In, select feature set from from the aspect of two, one is whether feature dissipates, and two is feature and target
Dependency.The present invention is in Filter（Filter）、Wrapper（Packaging）And Embedded（It is embedded）These three feature selection theory sides
Under the guidance of method, using selecting feature set in following engineering：1）Calculated using the X 2 test of Filter, mutual information method each
Individual feature and the dependency of response variable, select feature according to relevance ranking；2）Logic recursion elimination using Wrapper is special
Method is levied, many wheel training are carried out using a basic mode type, after often wheel training, the feature of some weight coefficients is eliminated, then based on new
Feature set carry out next round training；3）Using the Method for Feature Selection based on penalty term of Embedded, using with penalty term
Basic mode type, in addition to filtering out feature, while being also carried out dimensionality reduction.
Step of the present invention（S22）The engineering method of middle characteristic processing includes following process step：（S221）By discretization
Method carries out discrete processes to multiple continuouss features；（S222）Nondimensionalization is carried out to continuouss feature by method for normalizing
Process, to solve the comparable sex chromosome mosaicism between data；（S223）Qualitative features are converted to calmly by the mode that mute coding is usually used
The impact of measure feature, such as sex, occupation to purchase financial product, this " quantization " is completed typically by " dummy variable " is introduced
's；Mute coding is a kind of mode for realizing One-Hot, it is possible to increase the precision of model.Wherein, step（S221）Middle discrete processes
When, age bracket, assets rank, monthly income rank client etc. are characterized in that special by continuouss such as age, total assets, monthly incomes
Levying carries out discretization and obtains；Sliding-model control has following benefit：1）More meet business visual angle；2）Sparse vector inner product multiplying
Speed is fast, and result of calculation is convenient to be stored, and easily extends；3）Feature after discretization has very strong robustness to abnormal data；4）
Facilitate the introduction of cross feature.Step（S222）In method for normalizing include min-max methods and z-score methods.min-max
Method is to carry out linear transformation to initial data, is mapped as the value of interval [0,1], and formula is：New data=(former data-minimum
Value)/(maximum-minimum)；Z-score methods are the standards that the average and standard deviation based on initial data carries out data
Change, on condition that initial data meets normal distribution, formula is：New data=(former data-average)/standard deviation.
Senior general is crossed for eigenmatrix causes the problems such as computationally intensive, the training time is long, and reduction eigenmatrix dimension is must
Indispensable.Step of the present invention（S24）Using based on the Method for Feature Selection of L1 penalty terms, PCA（PCA）With it is linear
Discriminant analysiss（LDA）Carry out Feature Dimension Reduction.PCA is a kind of unsupervised dimension reduction method, and the sample allowed after mapping has maximum
Diversity；LDA is a kind of dimension reduction method for having supervision, and the sample allowed after mapping has best classification performance.Through Feature Engineering
Afterwards, final choice：Customer ID, sex, length of service, age bracket, marital status, educational background, occupation, the time limit of opening an account, whether silver
Office staff's work, personal asset rank, personal liability rank, client's acceptable level of risk are other, fund generic attribute（Part）, financing generic
Property（Part）, precious metal attribute（Part）, insurance generic attribute（Part）Deng 20 multinomial features, and form feature database and be supplied to mould
Type training is used.
Step of the present invention（S03）Naive Bayes Classifier calculating and assessment are carried out using data set, and constantly adjusts visitor
Family feature set and parameter, obtain the process of model optimized parameter through successive ignition, and which comprises the steps：（S31）According to equal
Even Distribution Principles, extract 80% data as training set from feature database at random, use for training module；It is left in feature database
20% data are used for model evaluation module as checking collection；（S32）Input training set, calculates and preserves each finance and produce
Product sell probability, every kind of feature correspondence each financial product conditional probability, in conjunction with parameter（Such as feature
The parameters such as weight）Set up the financial product recommended models storehouse based on Naive Bayes Classification；（S33）Input validation collection, to client
Purchase financial product is predicted, and concentrates the financial product of actual purchase to calculate hit rate, statistics according to predicting the outcome and verifying
Go out evaluation metricses and carry out assessment models.
Step（S33）In, judge whether evaluation metricses exceed index threshold values, if evaluation metricses are less than threshold values, adjust
The feature set and parameter of whole input（The such as parameter such as feature weight）, re-start a new round data set extract, model training and
Model evaluation process.Through successive ignition, until evaluation metricses exceed threshold values, model training and assessment are just accused and are completed.Finally,
Take repeatedly qualified evaluation metricses and calculate meansigma methodss, as the desired value of model evaluation.
Naive Bayes Classification Algorithm is as follows in the utilization principle of the present invention：
If, { sex, length of service, age bracket, marital status, educational background, occupation, open an account such as x=
The time limit, if bank employee, personal asset rank ... }, x is a client characteristics collection to be sorted, and each a is the one of client x
Individual characteristic attribute.Financial product set is：, such as C={ fund product 1, fund product 2, fund product
3, finance product 1, finance product 2, finance product 3, noble metal products 1, noble metal products 2, noble metal products 3,
…}。
Calculate the probability that client buys each financial product：。
If, thenThe finance of client's purchase is best suitable for exactly
Product.
The step of most critical is to calculate the probability that client buys each financial product。
According to naive Bayesian theorem,, denominator P (x) is constant for all financial products, is calculated
And compare, maximum is correspondingThe financial product of client's purchase is best suitable for exactly.
Assume that each characteristic attribute is conditional sampling, then
During model training, calculate and preserve each financial product sells probabilityEach finance corresponding with every kind of feature is produced
The conditional probability of product
Step of the present invention（S04）It is, with the financial product recommended models storehouse based on Naive Bayes Classification, visitor to be carried out to client
The matching process of family feature and financial product, which comprises the steps：（S41）To the client's body sent using AM access module
Part card number, bank's card number, fingerprint image, face-image carry out client's identification, obtain client characteristics collection；（S42）According to client's
Feature set reads the conditional probability of each financial product of every kind of feature correspondence, selling in conjunction with each financial product in model library
Sell probability and parameter calculates financial product classification and ordination value；Corresponding financial product is predicted according to financial product classification and ordination value
Collection；（S43）Screened according to financial product collection of the screening rule to prediction out, obtain final financial product, and pass through
Client is pushed to using AM access module.Step（S41）In, if the new client of bank, then set in advance to lead referral
Financial product；If the storage client of bank, then obtain the feature set of the client from model library, and lead to
Crossing financial product forecast function carries out matching for client characteristics and financial product.Step（S43）In, screening rule will according to business
Asking carries out Parametric Definition, and parameter includes：Financial product category preferences, financial product priority, the finance product of default recommendation
Product, portfolio parameter etc..
Step of the present invention（S05）Client of the middle financial product real time propelling movement that model prediction module is matched to application end
When, the push mode for using includes Web Service, HTML5, JavaScript, REST and Socket, and which supports site row
The financial product of team's machine, site self-service equipment, wifi, Web bank, Mobile banking APP and handset Wechat Bank application is recommended to connect
Enter function.
Invention additionally discloses a kind of bank finance product real-time recommendation system based on Naive Bayes Classification, which includes number
According to acquisition module, Feature Engineering module, model training module, model prediction module, using AM access module and data repository；Number
Include basic database, feature database and model library according to thesauruss；Data acquisition module is used to gather banking system, periphery system
System, the data of the Internet/mobile Internet, and process and store basic database；Feature Engineering module is for full dose visitor
The data attribute at family carries out characteristic processing, is converted to data characteristicses and forms feature database；Model training module is for by feature database
Data be divided into training dataset and checking data set, be respectively used to the training and assessment of Naive Bayes Classification Algorithm model,
Jing successive ignitions obtain model library；Model prediction module is used to recognize client characteristics, and combines Naive Bayes Classification Algorithm mould
The financial product that type storehouse real-time matching client most possibly buys；Using AM access module be used for obtain client identification card number or
Bank's card number simultaneously submits to model prediction module, then the financial product real time propelling movement that model prediction module is matched is to application end
Client.The present invention adopts distributed file system（Such as the HDFS of Hadoop）Storage basic database and feature database, using pass
It is type data base（Such as MySQL）Preservation model storehouse；Using distributed memory computing technique（Such as Spark）Carry out data prediction,
Feature Engineering, model training and model prediction.
The present invention is built upon big data platform based on the bank finance product real-time recommendation system of Naive Bayes Classification
A kind of intelligent data product on the basis of technology and algorithm engine, real time individual precisely recommend to be belonged to according to the population of user
Property, hobby and purchasing behavior, the information for recommending client interested to user in real time and commodity, to help bank as its client
Completely personalized decision support and information service are provided.
Data collecting module collected banking system of the present invention, peripheral system, client's number of the Internet/mobile Internet
According to, and process and store basic database；Feature Engineering module carries out characteristic processing to the data attribute of full dose client, conversion
Feature database is formed for data characteristicses；Feature database data is divided into training dataset and checking data set by model training module, respectively
For the training and assessment of Naive Bayes Classification Algorithm model, Jing successive ignitions obtain model library；System is accessed by application
Module obtains the identification card number or bank's card number of client, model prediction module identification client characteristics, and combines naive Bayesian
The financial product that sorting algorithm model library real-time matching client most possibly buys, then by giving using AM access module real time propelling movement
The client of application end；The transaction record of the financial product that client's purchase is recommended feeds back to the system by data acquisition module and is formed
Closed loop, carries out continuous updating and optimization to model.
Data acquisition module of the present invention includes data preprocessing module, and data preprocessing module is for the data to collecting
Carry out quality analysiss, data scrubbing and Data Integration；And the data of data collecting module collected include client properties data, finance
Product attribute data and customer action data；And data acquisition module support data acquisition modes include offline acquisition mode and
Online acquisition mode；The Data Source of data collecting module collected includes：1）The customer information of kernel business of bank system, deposit loan
The data such as money, the credit card, fund, financing and insurance；2）The wage of local characteristic service system, charges for water and electricity, management for infrastructure fee and
The transaction data such as communication expense；3）People's row collage-credit data of peripheral system；4）The Web bank of the Internet/mobile Internet, wechat
The data such as bank, mobile banking APP, social media and electric business.
Feature of present invention engineering module includes feature candidate block, feature processing block, feature selection module and feature drop
Dimension module；Feature candidate block selects suitable data attribute as candidate feature for being analyzed to all client properties,
Simultaneously candidate feature or various data splittings are calculated, new feature is derived；Feature processing block is for candidate
Feature carries out characteristic processing；Feature selection module is used for the feature set for selecting to obtain model and algorithm top performance；Feature Dimension Reduction
Module is used to reduce eigenmatrix dimension, and forms feature database and be supplied to model training module to use.
Wherein, feature processing block includes descretization module, normalization module and mute coding module；Descretization module is used for
Discrete processes are carried out to multiple continuouss features；Normalization module is for carrying out nondimensionalization process to continuouss feature；Mute volume
Code module is for being converted to quantitative characteristic by qualitative features.
Wherein, model training module includes data set extraction module, training module and model evaluation module；Data set is extracted
Module is used for the random data from feature database extraction 80% as training set, uses for training module；It is left 20% in feature database
Data as checking collection, for model evaluation module use；Training module is used to be input into training set, calculates and preserve each gold
Melt the conditional probability for selling probability each financial product corresponding with every kind of feature of product, and set up based on Piao for incorporating parametric
The financial product recommended models storehouse of plain Bayes's classification；Model evaluation module is used for input validation collection, and client's purchase finance is produced
Product are predicted, and concentrate the financial product of actual purchase to calculate hit rate according to predicting the outcome and verifying, count evaluation metricses
Carry out assessment models.
Wherein, model prediction module includes client identification module, financial product prediction module and financial product screening module；
Client identification module is for the customer's identity card number to sending using AM access module, bank's card number, fingerprint image, face figure
As being identified；Financial product prediction module corresponds to each for every kind of feature is read in model library according to the feature set of client
The conditional probability of financial product, calculates financial product classification and ordination value in conjunction with sell probability and the parameter of each financial product；
Corresponding financial product collection is predicted according to financial product classification and ordination value；Financial product screening module is for predicting out
Financial product collection is screened, and obtains final financial product, and by being pushed to client using AM access module.
The present invention realizes load balancing using cluster using AM access module, and high concurrent, highly reliably response application API please
Ask, request is sent to model prediction module carries out client's financial product prediction, receive the recommendation results of return and be pushed to visitor
Family.The API access ways provided using AM access module include Web Service, HTML5, JavaScript, REST and Socket
Deng, and which supports site queue machine, site self-service equipment, wifi, Web bank, Mobile banking APP and handset Wechat bank etc.
Using financial product recommend access function.
The present invention, is examined based on Feature Engineering and model training and fully with Naive Bayes Classification Algorithm as core
The features such as data source is more, data volume is big, data mart modeling is complicated, financial product recommends requirement of real-time high are considered, have been directed to
The design of property and structure.
The present invention is changed conventional marketing personnel and is manually marketed etc. using phone or using manipulation type marketing system
Traditional Marketing pattern, is innovatively replaced artificial operation by Machine automated marketing flow process, is calculated by Naive Bayes Classification
Method engine replaces artificial judgment, realizes round-the-clock uninterrupted operation, full Products Show, full channel services client, solves " product letter
The problems such as breath overload " and " long-tail product ", help client to be quickly found out suitable product, improve customer experience, increase client's
Viscosity, improving product conversion ratio, the competitiveness for improving bank.
The present invention adopts distributed big data platform（Such as Hadoop）Realize smooth extension and the system of system expensive storage
The linear lifting of performance, solves, meet full dose client comprehensively and deposit
Storage and extension demand；Using big data distributed computing framework（Such as MapReduce）With distributed memory computing technique（Such as
Spark）Off-line data processing and the online performance for calculating in real time are improved greatly, satisfaction completes complete in reasonable time window
The processing of amount customer data and the requirement for calculating in real time.
The present invention realizes load balancing using cluster using AM access module, and high concurrent, highly reliably response application API please
Ask；Model prediction module adopts distributed memory computing technique（Such as Spark）Carry out calculating in real time and screening financial product online；
Therefore complete within 1 second from asking to return recommendation results whole process, realize the second level response to asking.
The present invention utilizes machine learning autonomic learning feedback result, implementation model self refresh, self-optimizing；Solve traditional meter
Calculate the problem that cannot be updated for a long time and optimize after model is reached the standard grade.
Presently preferred embodiments of the present invention is the foregoing is only, not to limit the present invention, all essences in the present invention
Within god and principle, any modification, equivalent substitution and improvements made etc. should be included within the scope of the present invention.
Claims (10)
1. a kind of real-time recommendation method of the bank finance product real-time recommendation system based on Naive Bayes Classification, its feature exist
In which comprises the steps：
（S01）Collection banking system, peripheral system, the data of the Internet/mobile Internet, and process and store basis
Data base；
（S02）Characteristic processing is carried out to the data attribute of full dose client, data characteristicses is converted to and is formed feature database；
（S03）Data in feature database are divided into into training dataset and checking data set, Naive Bayes Classification calculation is respectively used to
The training and assessment of method model, Jing successive ignitions obtain model library；
（S04）Identification client characteristics, and most possibly buy with reference to Naive Bayes Classification Algorithm model library real-time matching client
Financial product；
（S05）Client of the financial product real time propelling movement that model prediction module is matched to application end.
2. the bank finance product real-time recommendation method based on Naive Bayes Classification as claimed in claim 1, its feature exist
In step（S01）Comprise the steps：
（S11）Client properties data, financial product attribute data and visitor are gathered by offline acquisition mode and online acquisition mode
Family behavioral data；
（S12）To step（S11）In the data that collect carry out quality analysiss, data scrubbing and Data Integration.
3. the bank finance product real-time recommendation method based on Naive Bayes Classification as claimed in claim 2, its feature exist
In step（S11）The Data Source of middle collection includes：1）The customer information of kernel business of bank system, loans and deposits, the credit card,
Fund, financing and insurance data；2）The wage of local characteristic service system, charges for water and electricity, management for infrastructure fee and communication expense number of deals
According to；3）People's row collage-credit data of peripheral system；4）The Web bank of the Internet/mobile Internet, wechat bank, mobile banking
APP, social media and electric quotient data.
4. the bank finance product real-time recommendation method based on Naive Bayes Classification as claimed in claim 3, its feature exist
In step（S02）Comprise the steps：
（S21）All client properties are analyzed, and suitable data attribute are selected as candidate feature, while to candidate feature
Or various data splittings are calculated, new feature is derived；
（S22）According to characteristic and the requirement of Naive Bayes Classification Algorithm model, to step（S11）In candidate feature carry out spy
Levy process；
（S23）Select to obtain the feature set of model and algorithm top performance；
（S24）Reduce eigenmatrix dimension, and form feature database being supplied to model training module to use.
5. the bank finance product real-time recommendation method based on Naive Bayes Classification as claimed in claim 4, its feature exist
In step（S22）The engineering method of middle characteristic processing includes following process step：
（S221）Discrete processes are carried out to multiple continuouss features by discretization method；
（S222）Nondimensionalization process is carried out to continuouss feature by method for normalizing；
（S223）Qualitative features are converted to by quantitative characteristic by mute coded method.
6. the bank finance product real-time recommendation method based on Naive Bayes Classification as claimed in claim 5, its feature exist
In step（S222）In method for normalizing include min-max methods and z-score methods.
7. the bank finance product real-time recommendation method based on Naive Bayes Classification as claimed in claim 6, its feature exist
In step（S03）Comprise the steps：
（S31）The random data from feature database extraction 80% are used for training module as training set；It is left 20% in feature database
Data as checking collection, for model evaluation module use；
（S32）Input training set, probability each finance corresponding with every kind of feature of selling for calculating and preserving each financial product are produced
The conditional probability of product, and the financial product recommended models storehouse based on Naive Bayes Classification is set up for incorporating parametric；
（S33）Input validation collection, is predicted to client's purchase financial product, concentrates actual purchase according to predicting the outcome and verifying
Financial product calculate hit rate, count evaluation metricses and carry out assessment models.
8. the bank finance product real-time recommendation method based on Naive Bayes Classification as claimed in claim 7, its feature exist
In step（S04）Comprise the steps：
（S41）The customer's identity card number that sends using AM access module, bank's card number, fingerprint image, face-image are carried out
Client's identification, obtains client characteristics collection；
（S42）The conditional probability of each financial product of every kind of feature correspondence is read in model library according to the feature set of client, then
Financial product classification and ordination value is calculated with reference to sell probability and the parameter of each financial product；According to financial product classification and ordination value
Predict corresponding financial product collection；
（S43）Prediction financial product collection out is screened, final financial product is obtained, and by applying AM access module
It is pushed to client.
9. the bank finance product real-time recommendation method based on Naive Bayes Classification as claimed in claim 8, its feature exist
In step（S05）The middle financial product real time propelling movement that model prediction module is matched to application end client when, what is used pushes away
Send mode to include Web Service, HTML5, JavaScript, REST and Socket, and which supports site queue machine, site certainly
The financial product of equipment, wifi, Web bank, Mobile banking APP and handset Wechat Bank application is helped to recommend access function.
10. a kind of bank finance product reality realized as claimed in any one of claims 1-9 wherein based on Naive Bayes Classification
When recommend method real-time recommendation system, it is characterised in which includes data acquisition module, Feature Engineering module, model training
Module, model prediction module, using AM access module and data repository；Data repository include basic database, feature database and
Model library；
Data acquisition module is used to gather banking system, peripheral system, the data of the Internet/mobile Internet, and processes
With storage to basic database；
Feature Engineering module carries out characteristic processing for the data attribute to full dose client, is converted to data characteristicses and forms feature
Storehouse；
Model training module is respectively used to simple shellfish for the data in feature database are divided into training dataset and checking data set
The training and assessment of leaf this sorting algorithm model, Jing successive ignitions obtain model library；
Model prediction module is used to recognize client characteristics, and combines Naive Bayes Classification Algorithm model library real-time matching client most
It is possible to the financial product bought；
It is used to obtain the identification card number or bank's card number of client and submit to model prediction module using AM access module, then by mould
Client of the financial product real time propelling movement that type prediction module is matched to application end.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611129818.9A CN106600369A (en) | 2016-12-09 | 2016-12-09 | Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611129818.9A CN106600369A (en) | 2016-12-09 | 2016-12-09 | Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106600369A true CN106600369A (en) | 2017-04-26 |
Family
ID=58597888
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611129818.9A Pending CN106600369A (en) | 2016-12-09 | 2016-12-09 | Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106600369A (en) |
Cited By (61)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107123053A (en) * | 2017-06-26 | 2017-09-01 | 杭州七桥金融信息服务有限公司 | A kind of method and system based on customer profile data Auto-matching financial product |
| CN107133836A (en) * | 2017-03-22 | 2017-09-05 | 无锡中科富农物联科技有限公司 | A kind of adaptive weighting combined recommendation algorithm |
| CN107220892A (en) * | 2017-05-27 | 2017-09-29 | 国家计算机网络与信息安全管理中心 | One kind melts data intelligence pretreating tool and method applied to magnanimity P2P net monetary allowances |
| CN107491533A (en) * | 2017-08-22 | 2017-12-19 | 安徽简道科技有限公司 | Security information individuation service system based on big data |
| CN107507068A (en) * | 2017-09-02 | 2017-12-22 | 广东奡风科技股份有限公司 | A kind of financial product real-time recommendation method based on random forests algorithm |
| CN107679946A (en) * | 2017-09-28 | 2018-02-09 | 平安科技（深圳）有限公司 | Fund Products Show method, apparatus, terminal device and storage medium |
| CN107730389A (en) * | 2017-09-30 | 2018-02-23 | 平安科技（深圳）有限公司 | Electronic installation, insurance products recommend method and computer-readable recording medium |
| CN107729540A (en) * | 2017-10-31 | 2018-02-23 | 努比亚技术有限公司 | Method, equipment and the computer of photo classification can storage mediums |
| CN107730372A (en) * | 2017-11-29 | 2018-02-23 | 济南浪潮高新科技投资发展有限公司 | The method and apparatus that a kind of Bank Hall personalized product based on recognition of face is recommended |
| CN107766500A (en) * | 2017-10-19 | 2018-03-06 | 国网冀北电力有限公司 | The auditing method of fixed assets card |
| CN107871269A (en) * | 2017-11-08 | 2018-04-03 | 广东奡风科技股份有限公司 | A kind of KNN of client's factor based on financial product Interactive matrix recommends method |
| CN107909473A (en) * | 2017-12-27 | 2018-04-13 | 中国银行股份有限公司 | A kind of Web bank's marketing method and device based on user behavior analysis |
| CN108182633A (en) * | 2018-01-30 | 2018-06-19 | 平安普惠企业管理有限公司 | Loan data processing method, device, computer equipment and storage medium |
| CN108399565A (en) * | 2017-10-09 | 2018-08-14 | 平安科技（深圳）有限公司 | Financial product recommendation apparatus, method and computer readable storage medium |
| CN108446986A (en) * | 2018-03-22 | 2018-08-24 | 张家林 | The data processing and computational methods of investment combination, device and equipment |
| CN108492190A (en) * | 2018-03-22 | 2018-09-04 | 张家林 | Intelligent security Portfolio Selection Based method, apparatus and equipment |
| CN108846687A (en) * | 2018-04-02 | 2018-11-20 | 平安科技（深圳）有限公司 | Client segmentation method, apparatus and 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 |
| CN109255638A (en) * | 2017-07-13 | 2019-01-22 | 北京融和友信科技股份有限公司 | A kind of mathematical model for excavating potential customers |
| CN109300050A (en) * | 2018-08-31 | 2019-02-01 | 平安科技（深圳）有限公司 | Insurance method for pushing, device and storage medium based on user's portrait |
| CN109300039A (en) * | 2018-12-05 | 2019-02-01 | 山东省城市商业银行合作联盟有限公司 | The method and system of intellectual product recommendation are carried out based on artificial intelligence and big data |
| CN109377388A (en) * | 2018-09-13 | 2019-02-22 | 平安医疗健康管理股份有限公司 | Medical insurance is insured method, apparatus, computer equipment and storage medium |
| CN109410075A (en) * | 2018-10-23 | 2019-03-01 | 广州市勤思网络科技有限公司 | Intelligence insurance recommended method and system based on Bayes |
| WO2019041520A1 (en) * | 2017-08-31 | 2019-03-07 | 平安科技（深圳）有限公司 | Social data-based method of recommending financial product, electronic device and medium |
| CN109447728A (en) * | 2018-09-07 | 2019-03-08 | 平安科技（深圳）有限公司 | Financial product recommended method, device, computer equipment and storage medium |
| CN109472646A (en) * | 2018-11-16 | 2019-03-15 | 广发证券股份有限公司 | A kind of financial product recommended method and device |
| CN109509082A (en) * | 2018-10-31 | 2019-03-22 | 中国银行股份有限公司 | The monitoring method and device of bank application system |
| WO2019061990A1 (en) * | 2017-09-30 | 2019-04-04 | 平安科技（深圳）有限公司 | User intention prediction method, electronic device, and computer readable storage medium |
| CN109614982A (en) * | 2018-10-18 | 2019-04-12 | 平安科技（深圳）有限公司 | Product analysis method, apparatus, computer equipment and storage medium |
| CN109658230A (en) * | 2018-12-11 | 2019-04-19 | 中盟盛世投资控股(深圳)有限公司 | Borrow or lend money processing method and processing device in internet |
| CN109858957A (en) * | 2019-01-17 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Finance product recommended method, device, computer equipment and storage medium |
| CN109871484A (en) * | 2019-01-31 | 2019-06-11 | 广州工程技术职业学院 | A kind of financial product real-time recommendation method |
| CN109903140A (en) * | 2019-03-07 | 2019-06-18 | 阿里巴巴集团控股有限公司 | A kind of credit services recommended method, device and equipment |
| CN109934512A (en) * | 2019-03-28 | 2019-06-25 | 努比亚技术有限公司 | A kind of training method and system of prediction model |
| CN110189191A (en) * | 2019-05-05 | 2019-08-30 | 苏宁易购集团股份有限公司 | A kind of financial product recommended method and system based on real-time model |
| CN110276668A (en) * | 2019-07-01 | 2019-09-24 | 中国工商银行股份有限公司 | The method and system that finance product intelligently pushing, matching degree determine |
| CN110335057A (en) * | 2019-04-30 | 2019-10-15 | 广发证券股份有限公司 | A kind of fund Precision Marketing Method that machine learning is merged with artificial rule |
| CN110348958A (en) * | 2019-06-28 | 2019-10-18 | 中信百信银行股份有限公司 | A kind of personalized recommendation method and system |
| CN110415084A (en) * | 2019-07-30 | 2019-11-05 | 中国工商银行股份有限公司 | A kind of product intelligent recommended method and device |
| CN110517151A (en) * | 2019-08-29 | 2019-11-29 | 中国农业银行股份有限公司 | A kind of Products Show method and system |
| CN110580648A (en) * | 2018-06-07 | 2019-12-17 | 上海诺亚投资管理有限公司 | financial information processing method and device based on artificial intelligence |
| CN110599228A (en) * | 2018-06-13 | 2019-12-20 | 北京智慧图科技有限责任公司 | Shop identification method |
| CN110781174A (en) * | 2019-10-15 | 2020-02-11 | 支付宝(杭州)信息技术有限公司 | Feature engineering modeling method and system using pca and feature intersection |
| US20200090269A1 (en) * | 2017-05-27 | 2020-03-19 | Alibaba Group Holding Limited | Data collection method and apparatus for risk evaluation, and electronic device |
| CN110956503A (en) * | 2019-11-29 | 2020-04-03 | 上海观安信息技术股份有限公司 | User identification method and device with loan demand based on user network behavior |
| CN111127077A (en) * | 2019-11-29 | 2020-05-08 | 中国建设银行股份有限公司 | Recommendation method and device based on stream computing |
| CN111159530A (en) * | 2019-11-19 | 2020-05-15 | 中国建设银行股份有限公司四川省分行 | System, apparatus, method and storage medium for automatically pushing financial product message |
| CN111488138A (en) * | 2020-04-10 | 2020-08-04 | 杭州顺藤网络科技有限公司 | B2B recommendation engine based on Bayesian algorithm and cosine algorithm |
| CN111506908A (en) * | 2020-04-10 | 2020-08-07 | 深圳新致软件有限公司 | Big data recommendation method, system and equipment for insurance industry |
| CN111951044A (en) * | 2020-07-30 | 2020-11-17 | 中国工商银行股份有限公司 | Bank terminal interaction method and system |
| CN112231550A (en) * | 2020-09-11 | 2021-01-15 | 重庆誉存大数据科技有限公司 | Credit financial product recommendation processing method and device |
| CN112580685A (en) * | 2020-11-19 | 2021-03-30 | 青岛檬豆网络科技有限公司 | MLCC (multilayer ceramic capacitor) capacitance parameter matching method |
| CN112950359A (en) * | 2021-03-30 | 2021-06-11 | 建信金融科技有限责任公司 | User identification method and device |
| CN113129154A (en) * | 2021-05-18 | 2021-07-16 | 中国银行股份有限公司 | Inter-row recommendation method and device for fund product information |
| CN113222734A (en) * | 2021-05-21 | 2021-08-06 | 中国银行股份有限公司 | Bank financial information recommendation system and method |
| CN113706266A (en) * | 2021-09-07 | 2021-11-26 | 中国银行股份有限公司 | Business recommendation method and system based on bank invoice data |
| CN113935776A (en) * | 2021-10-19 | 2022-01-14 | 中国银行股份有限公司 | Accurate marketing method, device and equipment based on big data |
| TWI769385B (en) * | 2019-07-03 | 2022-07-01 | 第一商業銀行股份有限公司 | Method and system for screening potential purchasers of financial products |
| US11403197B2 (en) | 2017-09-30 | 2022-08-02 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method and device for controlling application, storage medium, and electronic device |
| CN114926287A (en) * | 2022-06-02 | 2022-08-19 | 河北源达信息技术股份有限公司 | Method and device for intelligently recommending investment products and electronic equipment |
| CN117333275A (en) * | 2023-10-12 | 2024-01-02 | 中国农业银行股份有限公司江苏省分行 | Marketing platform construction method based on big data |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104851026A (en) * | 2015-05-21 | 2015-08-19 | 上海宾谷网络科技有限公司 | Big data based bid native advertisement reward system for positioning user in real time, and method |
| CN104866626A (en) * | 2015-06-15 | 2015-08-26 | ***通信集团黑龙江有限公司 | Method and device for recommending telecommunication service |
| CN105528374A (en) * | 2014-10-21 | 2016-04-27 | 苏宁云商集团股份有限公司 | A commodity recommendation method in electronic commerce and a system using the same |
| CN106021379A (en) * | 2016-05-12 | 2016-10-12 | 深圳大学 | Personalized recommendation method and system based on user preference |
-
2016
- 2016-12-09 CN CN201611129818.9A patent/CN106600369A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105528374A (en) * | 2014-10-21 | 2016-04-27 | 苏宁云商集团股份有限公司 | A commodity recommendation method in electronic commerce and a system using the same |
| CN104851026A (en) * | 2015-05-21 | 2015-08-19 | 上海宾谷网络科技有限公司 | Big data based bid native advertisement reward system for positioning user in real time, and method |
| CN104866626A (en) * | 2015-06-15 | 2015-08-26 | ***通信集团黑龙江有限公司 | Method and device for recommending telecommunication service |
| CN106021379A (en) * | 2016-05-12 | 2016-10-12 | 深圳大学 | Personalized recommendation method and system based on user preference |
Cited By (79)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107133836A (en) * | 2017-03-22 | 2017-09-05 | 无锡中科富农物联科技有限公司 | A kind of adaptive weighting combined recommendation algorithm |
| US20200090269A1 (en) * | 2017-05-27 | 2020-03-19 | Alibaba Group Holding Limited | Data collection method and apparatus for risk evaluation, and electronic device |
| CN107220892A (en) * | 2017-05-27 | 2017-09-29 | 国家计算机网络与信息安全管理中心 | One kind melts data intelligence pretreating tool and method applied to magnanimity P2P net monetary allowances |
| CN107220892B (en) * | 2017-05-27 | 2020-10-16 | 国家计算机网络与信息安全管理中心 | Intelligent preprocessing tool and method applied to massive P2P network loan financial data |
| CN107123053A (en) * | 2017-06-26 | 2017-09-01 | 杭州七桥金融信息服务有限公司 | A kind of method and system based on customer profile data Auto-matching financial product |
| CN109255638A (en) * | 2017-07-13 | 2019-01-22 | 北京融和友信科技股份有限公司 | A kind of mathematical model for excavating potential customers |
| CN107491533A (en) * | 2017-08-22 | 2017-12-19 | 安徽简道科技有限公司 | Security information individuation service system based on big data |
| WO2019041520A1 (en) * | 2017-08-31 | 2019-03-07 | 平安科技（深圳）有限公司 | Social data-based method of recommending financial product, electronic device and medium |
| CN107507068A (en) * | 2017-09-02 | 2017-12-22 | 广东奡风科技股份有限公司 | A kind of financial product real-time recommendation method based on random forests algorithm |
| CN107679946B (en) * | 2017-09-28 | 2021-09-10 | 平安科技（深圳）有限公司 | Fund product recommendation method and device, terminal equipment and storage medium |
| CN107679946A (en) * | 2017-09-28 | 2018-02-09 | 平安科技（深圳）有限公司 | Fund Products Show method, apparatus, terminal device and storage medium |
| US11403197B2 (en) | 2017-09-30 | 2022-08-02 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method and device for controlling application, storage medium, and electronic device |
| CN107730389A (en) * | 2017-09-30 | 2018-02-23 | 平安科技（深圳）有限公司 | Electronic installation, insurance products recommend method and computer-readable recording medium |
| WO2019061990A1 (en) * | 2017-09-30 | 2019-04-04 | 平安科技（深圳）有限公司 | User intention prediction method, electronic device, and computer readable storage medium |
| CN108399565A (en) * | 2017-10-09 | 2018-08-14 | 平安科技（深圳）有限公司 | Financial product recommendation apparatus, method and computer readable storage medium |
| CN107766500A (en) * | 2017-10-19 | 2018-03-06 | 国网冀北电力有限公司 | The auditing method of fixed assets card |
| CN107729540A (en) * | 2017-10-31 | 2018-02-23 | 努比亚技术有限公司 | Method, equipment and the computer of photo classification can storage mediums |
| CN107729540B (en) * | 2017-10-31 | 2021-04-16 | 努比亚技术有限公司 | Method, apparatus and computer-readable storage medium for photo classification |
| CN107871269A (en) * | 2017-11-08 | 2018-04-03 | 广东奡风科技股份有限公司 | A kind of KNN of client's factor based on financial product Interactive matrix recommends method |
| CN107730372A (en) * | 2017-11-29 | 2018-02-23 | 济南浪潮高新科技投资发展有限公司 | The method and apparatus that a kind of Bank Hall personalized product based on recognition of face is recommended |
| CN107909473A (en) * | 2017-12-27 | 2018-04-13 | 中国银行股份有限公司 | A kind of Web bank's marketing method and device based on user behavior analysis |
| CN108182633A (en) * | 2018-01-30 | 2018-06-19 | 平安普惠企业管理有限公司 | Loan data processing method, device, computer equipment and storage medium |
| CN108182633B (en) * | 2018-01-30 | 2020-08-04 | 平安普惠企业管理有限公司 | Loan data processing method, loan data processing device, loan data processing program, and computer device and storage medium |
| CN108492190A (en) * | 2018-03-22 | 2018-09-04 | 张家林 | Intelligent security Portfolio Selection Based method, apparatus and equipment |
| CN108446986A (en) * | 2018-03-22 | 2018-08-24 | 张家林 | The data processing and computational methods of investment combination, device and equipment |
| WO2019192130A1 (en) * | 2018-04-02 | 2019-10-10 | 平安科技（深圳）有限公司 | Customer classification method and device and storage medium |
| CN108846687A (en) * | 2018-04-02 | 2018-11-20 | 平安科技（深圳）有限公司 | Client segmentation method, apparatus and storage medium |
| CN110580648A (en) * | 2018-06-07 | 2019-12-17 | 上海诺亚投资管理有限公司 | financial information processing method and device based on artificial intelligence |
| CN110599228A (en) * | 2018-06-13 | 2019-12-20 | 北京智慧图科技有限责任公司 | Shop identification method |
| CN110599228B (en) * | 2018-06-13 | 2022-08-23 | 北京智慧图科技有限责任公司 | Shop identification method |
| 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 |
| CN109300050A (en) * | 2018-08-31 | 2019-02-01 | 平安科技（深圳）有限公司 | Insurance method for pushing, device and storage medium based on user's portrait |
| CN109447728A (en) * | 2018-09-07 | 2019-03-08 | 平安科技（深圳）有限公司 | Financial product recommended method, device, computer equipment and storage medium |
| CN109377388A (en) * | 2018-09-13 | 2019-02-22 | 平安医疗健康管理股份有限公司 | Medical insurance is insured method, apparatus, computer equipment and storage medium |
| CN109377388B (en) * | 2018-09-13 | 2023-08-18 | 深圳平安医疗健康科技服务有限公司 | Medical insurance application method, medical insurance application device, computer equipment and storage medium |
| CN109614982A (en) * | 2018-10-18 | 2019-04-12 | 平安科技（深圳）有限公司 | Product analysis method, apparatus, computer equipment and storage medium |
| CN109410075A (en) * | 2018-10-23 | 2019-03-01 | 广州市勤思网络科技有限公司 | Intelligence insurance recommended method and system based on Bayes |
| CN109509082A (en) * | 2018-10-31 | 2019-03-22 | 中国银行股份有限公司 | The monitoring method and device of bank application system |
| CN109509082B (en) * | 2018-10-31 | 2022-02-25 | 中国银行股份有限公司 | Monitoring method and device for bank application system |
| CN109472646A (en) * | 2018-11-16 | 2019-03-15 | 广发证券股份有限公司 | A kind of financial product recommended method and device |
| CN109300039A (en) * | 2018-12-05 | 2019-02-01 | 山东省城市商业银行合作联盟有限公司 | The method and system of intellectual product recommendation are carried out based on artificial intelligence and big data |
| CN109658230A (en) * | 2018-12-11 | 2019-04-19 | 中盟盛世投资控股(深圳)有限公司 | Borrow or lend money processing method and processing device in internet |
| CN109858957A (en) * | 2019-01-17 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Finance product recommended method, device, computer equipment and storage medium |
| CN109871484A (en) * | 2019-01-31 | 2019-06-11 | 广州工程技术职业学院 | A kind of financial product real-time recommendation method |
| CN109871484B (en) * | 2019-01-31 | 2022-02-18 | 广州工程技术职业学院 | Real-time financial product recommendation method |
| CN109903140A (en) * | 2019-03-07 | 2019-06-18 | 阿里巴巴集团控股有限公司 | A kind of credit services recommended method, device and equipment |
| CN109934512B (en) * | 2019-03-28 | 2024-02-09 | 努比亚技术有限公司 | Prediction model training method and training system |
| CN109934512A (en) * | 2019-03-28 | 2019-06-25 | 努比亚技术有限公司 | A kind of training method and system of prediction model |
| CN110335057A (en) * | 2019-04-30 | 2019-10-15 | 广发证券股份有限公司 | A kind of fund Precision Marketing Method that machine learning is merged with artificial rule |
| CN110189191A (en) * | 2019-05-05 | 2019-08-30 | 苏宁易购集团股份有限公司 | A kind of financial product recommended method and system based on real-time model |
| CN110348958A (en) * | 2019-06-28 | 2019-10-18 | 中信百信银行股份有限公司 | A kind of personalized recommendation method and system |
| CN110276668A (en) * | 2019-07-01 | 2019-09-24 | 中国工商银行股份有限公司 | The method and system that finance product intelligently pushing, matching degree determine |
| TWI769385B (en) * | 2019-07-03 | 2022-07-01 | 第一商業銀行股份有限公司 | Method and system for screening potential purchasers of financial products |
| CN110415084A (en) * | 2019-07-30 | 2019-11-05 | 中国工商银行股份有限公司 | A kind of product intelligent recommended method and device |
| CN110415084B (en) * | 2019-07-30 | 2022-10-21 | 中国工商银行股份有限公司 | Intelligent product recommendation method and device |
| CN110517151A (en) * | 2019-08-29 | 2019-11-29 | 中国农业银行股份有限公司 | A kind of Products Show method and system |
| CN110781174A (en) * | 2019-10-15 | 2020-02-11 | 支付宝(杭州)信息技术有限公司 | Feature engineering modeling method and system using pca and feature intersection |
| CN111159530A (en) * | 2019-11-19 | 2020-05-15 | 中国建设银行股份有限公司四川省分行 | System, apparatus, method and storage medium for automatically pushing financial product message |
| CN111127077A (en) * | 2019-11-29 | 2020-05-08 | 中国建设银行股份有限公司 | Recommendation method and device based on stream computing |
| CN110956503B (en) * | 2019-11-29 | 2024-02-23 | 上海观安信息技术股份有限公司 | User identification method and device with lending requirements based on user network behaviors |
| CN110956503A (en) * | 2019-11-29 | 2020-04-03 | 上海观安信息技术股份有限公司 | User identification method and device with loan demand based on user network behavior |
| CN111488138A (en) * | 2020-04-10 | 2020-08-04 | 杭州顺藤网络科技有限公司 | B2B recommendation engine based on Bayesian algorithm and cosine algorithm |
| CN111506908A (en) * | 2020-04-10 | 2020-08-07 | 深圳新致软件有限公司 | Big data recommendation method, system and equipment for insurance industry |
| CN111488138B (en) * | 2020-04-10 | 2023-08-04 | 杭州顺藤网络科技有限公司 | B2B recommendation engine based on Bayesian algorithm and cosine algorithm |
| CN111506908B (en) * | 2020-04-10 | 2023-02-17 | 深圳新致软件有限公司 | Big data recommendation method, system and equipment for insurance industry |
| CN111951044A (en) * | 2020-07-30 | 2020-11-17 | 中国工商银行股份有限公司 | Bank terminal interaction method and system |
| CN112231550A (en) * | 2020-09-11 | 2021-01-15 | 重庆誉存大数据科技有限公司 | Credit financial product recommendation processing method and device |
| CN112231550B (en) * | 2020-09-11 | 2024-05-17 | 重庆誉存大数据科技有限公司 | Credit financial product recommendation processing method and device |
| CN112580685B (en) * | 2020-11-19 | 2022-10-28 | 青岛檬豆网络科技有限公司 | MLCC (multilayer ceramic capacitor) capacitor parameter matching method |
| CN112580685A (en) * | 2020-11-19 | 2021-03-30 | 青岛檬豆网络科技有限公司 | MLCC (multilayer ceramic capacitor) capacitance parameter matching method |
| CN112950359A (en) * | 2021-03-30 | 2021-06-11 | 建信金融科技有限责任公司 | User identification method and device |
| CN113129154A (en) * | 2021-05-18 | 2021-07-16 | 中国银行股份有限公司 | Inter-row recommendation method and device for fund product information |
| CN113222734A (en) * | 2021-05-21 | 2021-08-06 | 中国银行股份有限公司 | Bank financial information recommendation system and method |
| CN113706266A (en) * | 2021-09-07 | 2021-11-26 | 中国银行股份有限公司 | Business recommendation method and system based on bank invoice data |
| CN113706266B (en) * | 2021-09-07 | 2024-05-28 | 中国银行股份有限公司 | Business recommendation method and system based on bank invoice data |
| CN113935776A (en) * | 2021-10-19 | 2022-01-14 | 中国银行股份有限公司 | Accurate marketing method, device and equipment based on big data |
| CN114926287A (en) * | 2022-06-02 | 2022-08-19 | 河北源达信息技术股份有限公司 | Method and device for intelligently recommending investment products and electronic equipment |
| CN117333275A (en) * | 2023-10-12 | 2024-01-02 | 中国农业银行股份有限公司江苏省分行 | Marketing platform construction method based on big data |
| CN117333275B (en) * | 2023-10-12 | 2024-03-19 | 中国农业银行股份有限公司江苏省分行 | Marketing platform construction method based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106600369A (en) | Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification | |
| Sun et al. | Predicting credit card delinquencies: An application of deep neural networks | |
| TW530234B (en) | Methods and systems for efficiently sampling portfolios for optimal underwriting | |
| CN108648074A (en) | Loan valuation method, apparatus based on support vector machines and equipment | |
| CN111861174B (en) | Credit assessment method for user portrait | |
| Chen | Classifying credit ratings for Asian banks using integrating feature selection and the CPDA-based rough sets approach | |
| CN110599336B (en) | Financial product purchase prediction method and system | |
| Ala’raj et al. | A deep learning model for behavioural credit scoring in banks | |
| Kulkarni et al. | Advanced credit score calculation using social media and machine learning | |
| Shi et al. | A novel imbalanced data classification approach based on logistic regression and Fisher discriminant | |
| CN111882420A (en) | Generation method of response rate, marketing method, model training method and device | |
| CN111695938B (en) | Product pushing method and system | |
| Pai et al. | Using internet search trends and historical trading data for predicting stock markets by the least squares support vector regression model | |
| CN111144899A (en) | Method and device for identifying false transactions and electronic equipment | |
| Yuping et al. | New methods of customer segmentation and individual credit evaluation based on machine learning | |
| CN111667307B (en) | Method and device for predicting financial product sales volume | |
| TWM613536U (en) | Investment risk scoring system for fund commodities | |
| Leventhal | Predictive Analytics for Marketers: Using Data Mining for Business Advantage | |
| Suthanthiradevi et al. | Modelling a Behavioral scoring system for lending loans using twitter | |
| Kim | Deep Learning for Uncertainty Measurement | |
| Liu et al. | Research on the identification and prediction of default risk of online lending platform customers | |
| CN118071483A (en) | Method for constructing retail credit risk prediction model and personal credit business Scorepsi model | |
| CN116703579A (en) | Client transaction prediction method, device and server | |
| KHANSONG et al. | Customer service improvement based on payment behavior electricity bill analysis using data mining approaches | |
| Ahmadani et al. | Improving Creditworthiness Prediction Using Preprocessing Stages and Feature Selection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170426 |