CN109509040A - Predict modeling method, marketing method and the device of fund potential customers - Google Patents
Predict modeling method, marketing method and the device of fund potential customers Download PDFInfo
- Publication number
- CN109509040A CN109509040A CN201910005765.7A CN201910005765A CN109509040A CN 109509040 A CN109509040 A CN 109509040A CN 201910005765 A CN201910005765 A CN 201910005765A CN 109509040 A CN109509040 A CN 109509040A
- Authority
- CN
- China
- Prior art keywords
- data
- fund
- potential customers
- model
- user
- 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
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- 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
- G06Q40/06—Asset management; Financial planning or analysis
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Human Resources & Organizations (AREA)
- Operations Research (AREA)
- Technology Law (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of modeling method, marketing method and device for predicting fund potential customers, the modeling method is comprising steps of building user information system;Using the user information system as foundation, prediction objective attribute target attribute Y is calculated, to predict that objective attribute target attribute Y and user information data as foundation, construct sample set D;Using the sample set D and machine learning algorithm as foundation, initial model is established, and optimize to the initial model, obtains the model of the prediction fund potential customers.The marketing method is to carry out financial product marketing using the model.Described device is for realizing the marketing method.It takes full advantage of the financial scenarios information such as financial market information, investment preference of the user to different type fund product of different investment styles under different financial market environment is captured using machine learning algorithm, to predict that the following month buys the client of certain class fund, be conducive to precisely commence business.
Description
Technical field
The present invention relates to modeling method, the marketing sides of financial market marketing domain, more particularly to prediction fund potential customers
Method and device.
Background technique
Under big data era background, major traditional industries all start to make the transition to " data driven type enterprise ".Enterprise wishes logical
It crosses data and finds business rule, customer demand is excavated, help business, which is optimized and promoted, brings direct valence to business
Value.From the point of view of strategic direction, enterprises decision-maker rule of thumb carries out decision and has greater risk, because behaviour will receive ring
The decision of some subjectivities is made in the influence of border and mood, and enterprise carries out verifying by data can be more objective.Finance, bond house
The business form of industry itself produces a large amount of quality data, has great tap value.How big data and people combined
Work intellectual technology provides reference and service for the decision and business development of related practitioner, becomes popular research direction instantly.State
Outer Switzerland, Huifeng and the Citibank is the forerunner of industry data digging technology application, each big bank and stock trader at home
In, the thought and technology of big data also gradually start to obtain practice and application in recent years.
There is great limitations for the Traditional Marketing mode of finance and money management product.Since a customer manager or investment care for
It asks and often corresponds to a large amount of service objects, and their professional standards are very different, this makes Traditional Marketing mode, and there is manpowers
At high cost, the problems such as client's coverage rate is low, counseling services subjectivity is strong.If can be by marketing process ' intelligence ', with Fund Type
Based on Automatic sieve select potential customers, this will not only can increase the business exchange hand of retail department, can more improve the satisfaction of client
Degree.The gradually concern by industry and academia of intelligent marketing mode, has formd more mature in e-commerce
' recommender system ' field.And in financial instrument industry, the concept of ' intelligence, which is thrown, to be cared for ' is still in the initial development stage.Safety science and technology
Real-time recommendation is carried out to client using the similarity of financial product historical trading behavior, Alibaba Co proposes that a kind of utilize produces
The method that incidence relation between product carries out finance product recommendation.Guang Dong Ao wind company discloses a kind of based on Naive Bayes Classification
Financial product real-time recommendation method, it is real-time in conjunction with Naive Bayes Classification Algorithm using the client characteristics chosen and handled well
Match the financial product that client most possibly buys.The Hong Kong and Shanghai Banking Corporation proposes one in primary financial big data Subject Scheming activity
Precision marketing scheme of the kind based on big data, utilizes the demand of the machine learning model insight into customer in its company's big data platform
With preference, sustainable marketing program is realized with this.
Although existing big data digging technology realizes the precision marketing of finance and money management product to a certain extent,
There is many Optimal improvements directions.Such as:
Data information utilization is not comprehensive.At present major part Generalization bounds primary concern is that client historical behavior data,
And the information such as the Assets of client, risk partiality are underused, and do not consider that current financial market situation determines to client
The influence of plan.
Summary of the invention
In order to solve the above-mentioned technical problem, the present invention proposes the modeling method of prediction fund potential customers a kind of, marketing side
Method and device, it has sufficiently used financial scenario information, and constructs the feature extracting method of financial scenario information, utilizes machine
Learning algorithm predicts the fund buying behavior of user, excavates all types of fund potential customers.
To achieve the goals above, the scheme that the present invention uses is:
A kind of modeling method for predicting fund potential customers, for constructing the model of prediction fund potential customers, including step
It is rapid:
Construct user information system;
Using the user information system as foundation, prediction objective attribute target attribute Y is calculated, to predict objective attribute target attribute Y and user information
Data are foundation, construct sample set D;
Using the sample set D and machine learning algorithm as foundation, initial model is established, and optimize to the initial model,
Obtain the model of the prediction fund potential customers.
The building user information system comprising steps of
Obtain initial data;
User information data are obtained in the initial data and are integrated, duplicate removal, storage, and user information body is obtained
System.
The initial data includes: business datum, outsourcing data and the financial business personnel of financial company are carrying out industry
The data collected when business and investigation.
The user information data include: customer transaction data, user tag data and reflection financial market market
Exponent data.
The user tag includes demographics label, social property label, Asset Tag, investment capacity label;It is described
Demographics label includes: user's gender, age, residence, education degree;The social property label includes: marriage situation,
Industry, income level are engaged in job site;The Asset Tag includes: total assets, initial assets, end of term assets;The investment
Ability label includes: the ability to ward off risks, customer grade.
The building sample sets D comprising steps of
Data set X is obtained, the data set X contains the data of Three Represents different types of information: customer transactional data,
The data of user tag data and reflection financial market market index;
It calculates prediction objective attribute target attribute Y and calculates each user each moon using the transaction record in the Transaction Information of user as foundation
Fund hold position accounting, accounting is held position using the fund of user's next month and the fund in this month hold position accounting difference as predicting target
Attribute Y;
Sample set D is constructed, the prediction objective attribute target attribute Y and data set X is whole by customer number and date progress
It closes, constructs sample set D.
Using the sample set D and machine learning algorithm as foundation, initial model is established, and optimize to the initial model and wrap
Include step:
Construct training set, verifying collection and test set;
It is optimized by optimization aim of hit rate, obtains the model of the prediction fund potential customers.
The machine learning algorithm includes: gradient boosting algorithm, random forests algorithm or SVM algorithm.
A kind of marketing method for predicting fund potential customers, comprising steps of
Using user information data as foundation, the model established using the method passes through channel APP push, mail, short message
Do the popularization of financial product.
A kind of marketing apparatus for predicting fund potential customers, comprising:
Model unit, for constructing the model for utilizing the method to establish;
Marketing unit, for using user information data as foundation, the model established using the method to pass through channel APP
The popularization of financial product is done in push, mail, short message.
The beneficial effects of the present invention are as follows: the financial scenarios information such as financial market information is taken full advantage of, using machine learning
Algorithm captures investment preference of the user to different type fund product of different investment styles under different financial market environment, thus
Prediction future buys the client of certain class fund month, is conducive to precisely commence business.
Detailed description of the invention
Fig. 1 shows the modeling method flow chart of the embodiment according to the present invention.
Fig. 2 shows the marketing method flow charts according to herein described embodiment.
Fig. 3 shows the feature extraction flow chart according to herein described embodiment.
Fig. 4, which is shown, establishes model and optimized flow chart according to herein described embodiment.
Specific embodiment
In order to be better understood by technical solution of the present invention, the invention will be further described by 1-4 with reference to the accompanying drawing.
As shown in Figure 1, a kind of modeling method for predicting fund potential customers, for constructing the mould of prediction fund potential customers
Type, comprising steps of
Obtain initial data;Initial data is obtained from multiple support channels, specific acquiring way is as shown in table 1.
Table 1
As shown in Table 1, the initial data includes: the business datum of financial company, passes through O&M, web log file data etc.
It obtains;Outsourcing data, financial company are obtained by purchase external data;Financial business personnel collection when commencing business and investigating
Data.
User information data are obtained in the initial data and are integrated, duplicate removal, storage, and user information body is constructed
System.The structural data and unstructured data obtained from initial data is integrated, and is extracted, and duplicate removal is stored into HDFS.
In order to construct perfect user information system, summarize the data of maximum magnitude as far as possible.It can be with subsidiary company system
Internal data and website crawler collect various user tags, mainly there is demographics label, including user's gender, year
Age, residence, education degree etc.;Social property label, including marriage situation, job site, are engaged in industry, income level etc.;
Asset Tag includes total assets, initial assets, end of term assets.Investment capacity label, including the ability to ward off risks, customer grade etc.,
User's portrait is constructed by these labels.According to different user tag building user's portraits, the essential information including user, money
Produce situation, investment capacity etc..The transformation in financial market also will affect the investment behavior of client, obtain from third party's data supplier
React the exponent data of market conditions.The index value in reflection financial market is extracted as characteristic attribute.Various types of data is integrated, is added
It is collected in hdfs after work statistics.
As shown in figure 3, extracting feature, sample set D is constructed: using the user information system as foundation, calculating prediction target
Attribute Y, to predict that objective attribute target attribute Y and user information data as foundation, construct sample set D.Specific steps include:
Obtain data set X.Data set X incorporates the data of Three Represents different types of information: customer transactional data, user
The data of representation data and reflection financial market information.Customer transactional data is with customer number, finance and money management product number, date
(year-month-day) is used as major key, has recorded client under different time to the hoard situation of different finance products.User's portrait number
According to a large amount of characteristic attributes for including reflection client's current asset situation, essential information, the investment capacity, mutual fund earnings situation, gold
Melt market information and has chosen Index of Shanghai Stock Exchange, Hu-Shen 300 index, Growth Enterprise Index, the middle full debt index of card, Nasdaq index.It is logical
It crosses the association attributes such as date, customer number and customer transactional data, user's representation data, index information data is spliced into original number
According to collection;Raw data set is integrated into the data set as unit of the moon, wherein initial data concentrates the attribute as unit of day logical
It crosses and calculates when the mode of month was processed into as unit of the moon, obtain data set X.
It constructs and calculates prediction objective attribute target attribute Y.The fund that transaction record based on client calculates each client each moon is held position
Accounting, using the fund of client's next month hold position accounting and this month fund hold position accounting difference as objective attribute target attribute Y, to represent
The fund of the client buys situation.The objective attribute target attribute Y is greater than 0 expression client and next month buys fund, otherwise indicates client
Next month does not buy fund.
As shown in figure 4, construction sample set D.By objective attribute target attribute Y and data set X by customer number and date (moon in year -) into
Row splicing (being exactly the join in SQL statement operation, record corresponding target data set X every and connect), constructs sample
This collection D.
Using the sample set D as foundation, building training set, verifying collection establish model with test set, with machine learning algorithm,
And processing is optimized to the model as optimization aim using hit rate and obtains static models, using the static models as foundation,
Time span is fixed as m months, is established as unit of the moon and rolls point, next month is carried out using newest input data pre-
It surveys, constructs roll modeling, obtain the model of the prediction fund potential customers.The roll modeling is to allow model to adapt to not
Same time model due to caused by turn of the market, customer priorities variation changes.It rolls dynamic model and mainly utilizes static mould
The optimal hyper parameter that type obtains when optimizing, the hyper parameter as static models.Therefore, the model of the prediction fund potential customers
Foundation can also construct again model by Optimal Parameters on the basis of rolling data collection.
Training set, verifying collection and test set are constructed on the basis of sample set D.Training set is T1 < t≤T2 in sample set D
Data, verifying integrate as the data of T2 < t≤T3 in sample set D, test set for T3 < t≤T4 in sample set D data, wherein
T1<T2<T3<T4.Wherein T1-T4 is a timing node data, for example T1 takes in January, 2017, and T2 takes in March, 2017, T3 T2
Some timing node data afterwards, T4 are some timing node data after T3.
Establish optimization aim.Using hit rate as model optimization target, hit rate are as follows:Predicted value is dropped
Sequence arrangement, ranking before choosing as desired as a result,Indicate predicted value, Y is true value, i.e.,
Adjustment parameter.After establishing good optimization aim, training set is imported to the machine learning mould established with machine learning algorithm
Among type, show that the hit rate of verifying collection, adjustment parameter improve the hit rate of verifying collection, obtain static models.
The machine learning algorithm has:
(1) gradient promotes tree-model
Gradient promotes its core of (gradient boosting) algorithm and is that, each tree is from the residual of previous all trees
Learn in difference, calculating each time for it is provided to reduce last residual error (residual), and residual in order to reduce these
Difference can establish a new tree-model on direction gradient (Gradient) of residual error reduction.Training set is put into algorithm mould
Type, the depth for adjusting the number of iterations and tree carry out Optimized model, promote the hit rate of verifying collection, and XGBoost algorithm also needs to adjust
Leaf node weight, specimen sample rate and penalty coefficient etc..
(2) random forest
Random forest (Random Forest) algorithm will carry out the sampling of row, column to the data of input.Row is sampled,
By the way of putting back to, so that when training, the sample that the input sample of every one tree is all not all of, relatively
It is not easy over-fitting occur.Column sampling is to select m (m < < M) features from M feature.The data after sampling are used later
The mode of fully nonlinear water wave establishes out decision tree, obtains final result by way of multiple decision tree ballots.If training data is defeated
Enter Random Forest model, adjustment parameter is number and the depth of tree to improve the hit rate of verifying collection.
(3) SVM (Support Vector Machine support vector machines)
Support vector machines can be widely used in statistical classification and regression analysis, and support vector machines belongs to generalized linear
The characteristics of classifier, this kind of classifier is that they can minimize experience error simultaneously and maximize Geometry edge area, therefore prop up
It holds vector machine and is also referred to as maximal margin area classifier.Objective function is
S.t., yi(wTxi+b)≥1-ξi, i=1-..., n
ξi>=0, i=1 ..., n
Wherein w and b is the characteristic parameter to be optimized, and may generally be considered as super classifying face, and y is that the class of training sample identifies,
ξiFor slack variable.Training sample is inputted into SVM model, adjustment parameter C (penalty coefficient of error sample) improves verifying collection
Hit rate.
It needs to change with financial market to cater to business, on the basis of static models, constructs roll modeling, obtain
Prediction fund potential customers' model.Time span is fixed as m (certainly by the method that rolling forecast method uses fixed length to roll
Definition) a month, it is established as unit of the moon and rolls point, using newest input data and optimized mould next month is predicted,
Machine learning model i (is denoted as Mi) using the point before i as training set, it is predicted using i point as test set, model Mi+mIt is to utilize i
Point before+m does training set, and i+m point is used as prediction, and so on.
After rolling forecast monthly compiles, enterprise can be carried out based on rolling forecast monthly based on these predictions
Analysis and Control work is arranged production based on rolling forecast operation and configuration resource first, it is actually complete secondly can to do last month
At the variance analysis of situation and the analysis of rolling forecast last month, reason of discrepancies is found out, the generation of this difference is compared based on yearly budget
Variance analysis is more meaningful because last month the prediction of next month is generated bigger difference just must detailed analysis reason, it is right
It acts in management very greatly, the forecasted variances of next month is analyzed in the prediction of lower next month and this month in addition to this it is possible to do last month,
It allows business department to explain the reason of discrepancies of this month prediction and Last Month's Forecast, is continuously improved the accuracy of prediction accordingly.
As shown in Fig. 2, a kind of marketing method for predicting fund potential customers, comprising steps of
Using user information data as foundation, the model established using the method, the result that model is obtained is applied to reality
In the business on border, pushed by channel APP, mail, short message marketing system do product promotion, it is variant that advertisement is launched.
A kind of marketing apparatus for predicting fund potential customers,
Model unit, for constructing the model for utilizing the method to establish;
Marketing unit, for using user information data as foundation, the model established using the method to pass through channel APP
The popularization of financial product is done in push, mail, short message.
The model unit includes:
Primary data module is obtained, for obtaining primary data;
Subscriber information storing module, in the initial data obtain user information data and integrated, duplicate removal,
Storage constructs user information system;
Characteristic extracting module calculates prediction objective attribute target attribute Y, using the user information system as foundation to predict target category
Property Y and user information data be foundation, construct sample set D;
Modeling module, using the sample set D as foundation, building training set, verifying collection are excellent with test set and with hit rate
Change target optimizes processing and obtains static models, and is established as unit of the moon on the basis of static models and roll point, building
Roll modeling obtains the model of the prediction fund potential customers.
The marketing apparatus further includes closing rule analysis and filter element: the cold list exported according to modeling module,
It filters out customer risk grade, client's list that investment time limit etc. is not consistent with fund product, guarantees the customer name preferably exported
Singly meet client's appropriateness specification of industry formulation.
The present invention is mainly used to predict the following client for buying certain class fund, is conducive to APP, determines in public platform and website
To push, it is also beneficial to business personnel and carries out precision marketing.
Above-mentioned, although the foregoing specific embodiments of the present invention is described with reference to the accompanying drawings, not protects model to the present invention
The limitation enclosed, those skilled in the art should understand that, based on the technical solutions of the present invention, those skilled in the art are not
Need to make the creative labor the various modifications or changes that can be made still within protection scope of the present invention.
Claims (10)
1. a kind of modeling method for predicting fund potential customers, for constructing the model of prediction fund potential customers, feature exists
In, comprising steps of
Construct user information system;
Using the user information system as foundation, prediction objective attribute target attribute Y is calculated, to predict objective attribute target attribute Y and user information data
For foundation, sample set D is constructed;
Using the sample set D and machine learning algorithm as foundation, initial model is established, and optimize to the initial model, obtained
The model of the prediction fund potential customers.
2. the modeling method of prediction fund potential customers according to claim 1, which is characterized in that the building user letter
Breath system comprising steps of
Obtain initial data;
User information data are obtained in the initial data and are integrated, duplicate removal, storage, and user information system is obtained.
3. a kind of modeling method for predicting fund potential customers according to claim 2, which is characterized in that the original number
The number collected according to the business datum, outsourcing data and financial business personnel that include: financial company when commencing business and investigating
According to.
4. a kind of modeling method for predicting fund potential customers according to claim 2, which is characterized in that user's letter
Breath data include: customer transaction data, user tag data and the exponent data for reflecting financial market market.
5. a kind of modeling method for predicting fund potential customers according to claim 4, which is characterized in that user's mark
Label include demographics label, social property label, Asset Tag, investment capacity label;The demographics label includes: use
Family gender, age, residence, education degree;The social property label includes: marriage situation, and industry is engaged in job site,
Income level;The Asset Tag includes: total assets, initial assets, end of term assets;The investment capacity label includes: wind resistance
Dangerous ability, customer grade.
6. a kind of modeling method for predicting fund potential customers according to claim 1, which is characterized in that the building sample
Product collection D comprising steps of
Data set X is obtained, the data set X contains the data of Three Represents different types of information: customer transactional data, user
The data of label data and reflection financial market market index;
It calculates prediction objective attribute target attribute Y and calculates the base of each user each moon using the transaction record in the Transaction Information of user as foundation
Gold is held position accounting, accounting is held position using the fund of user's next month and the fund in this month hold position accounting difference as predicting objective attribute target attribute
Y;
Sample set D is constructed, the prediction objective attribute target attribute Y and data set X is integrated by customer number and date, structure
Build sample set D.
7. a kind of modeling method for predicting fund potential customers according to claim 1, which is characterized in that with the sample
Integrate D and machine learning algorithm as foundation, establish initial model, and the initial model is optimized comprising steps of
Construct training set, verifying collection and test set;
It is optimized by optimization aim of hit rate, obtains the model of the prediction fund potential customers.
8. a kind of modeling method for predicting fund potential customers according to claim 7, which is characterized in that the engineering
Practising algorithm includes: gradient boosting algorithm, random forests algorithm or SVM algorithm.
9. a kind of marketing method for predicting fund potential customers, which is characterized in that comprising steps of
Using user information data as foundation, the model established using any the method for claim 1-8 is pushed away by channel APP
It send, mail, short message do the popularization of financial product.
10. a kind of marketing apparatus for predicting fund potential customers characterized by comprising
Model unit, for constructing the model for utilizing any the method for claim 1-8 to establish;
Marketing unit is used for using user information data as foundation, the model established using any the method for claim 1-8,
The popularization of financial product is done by channel APP push, mail, short message.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910005765.7A CN109509040A (en) | 2019-01-03 | 2019-01-03 | Predict modeling method, marketing method and the device of fund potential customers |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910005765.7A CN109509040A (en) | 2019-01-03 | 2019-01-03 | Predict modeling method, marketing method and the device of fund potential customers |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109509040A true CN109509040A (en) | 2019-03-22 |
Family
ID=65756222
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910005765.7A Pending CN109509040A (en) | 2019-01-03 | 2019-01-03 | Predict modeling method, marketing method and the device of fund potential customers |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109509040A (en) |
Cited By (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110147813A (en) * | 2019-04-04 | 2019-08-20 | 深圳价值在线信息科技股份有限公司 | A kind of user draws a portrait construction method, device, storage medium and server |
CN110175873A (en) * | 2019-05-24 | 2019-08-27 | 深圳市无忧宝科技有限公司 | Precision Marketing Method and device |
CN110210913A (en) * | 2019-06-14 | 2019-09-06 | 重庆邮电大学 | A kind of businessman frequent customer's prediction technique based on big data |
CN110223166A (en) * | 2019-06-14 | 2019-09-10 | 哈尔滨哈银消费金融有限责任公司 | The prediction technique and equipment of consumer finance user's overdue loan based on big data |
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 |
CN110427555A (en) * | 2019-07-27 | 2019-11-08 | 北京智数时空科技有限公司 | A kind of target user's needing forecasting method and system based on user behavior preference |
CN110458592A (en) * | 2019-06-18 | 2019-11-15 | 北京海致星图科技有限公司 | Knowledge based map and machine learning algorithm excavate the potential credit client method of bank |
CN110674100A (en) * | 2019-09-28 | 2020-01-10 | 上海浦东发展银行股份有限公司信用卡中心 | User demand prediction method and framework based on full-channel operation data |
CN110688553A (en) * | 2019-08-13 | 2020-01-14 | 平安科技(深圳)有限公司 | Information pushing method and device based on data analysis, computer equipment and storage medium |
CN111428092A (en) * | 2020-03-20 | 2020-07-17 | 北京中亦安图科技股份有限公司 | Accurate bank marketing method based on graph model |
CN111598580A (en) * | 2020-04-26 | 2020-08-28 | 杭州云象网络技术有限公司 | XGboost algorithm-based block chain product detection method, system and device |
CN111783818A (en) * | 2020-04-10 | 2020-10-16 | 南通大学 | Accurate marketing method based on Xgboost and DBSCAN |
CN112184292A (en) * | 2020-09-16 | 2021-01-05 | 中国农业银行股份有限公司河北省分行 | Marketing method and device based on artificial intelligence decision tree |
CN112256964A (en) * | 2020-10-22 | 2021-01-22 | 重庆邮电大学 | Financial institution potential customer recommendation method based on multi-dimensional data learning |
CN112308351A (en) * | 2019-07-26 | 2021-02-02 | 上海漕泾热电有限责任公司 | Heat supply balance prediction system and method for cogeneration power plant |
CN112329928A (en) * | 2020-12-30 | 2021-02-05 | 四川新网银行股份有限公司 | Heterogeneous model-based user satisfaction analysis method |
CN113313572A (en) * | 2021-05-28 | 2021-08-27 | 上海浦东发展银行股份有限公司 | Model identification method based on accumulation fund point-credit customer |
CN113421133A (en) * | 2021-07-21 | 2021-09-21 | 赛诺数据科技(南京)有限公司 | Network marketing system based on customer matching |
CN114004651A (en) * | 2021-11-02 | 2022-02-01 | 武汉市伟荣发信息科技有限公司 | Information statistics and arrangement system for customer-obtaining channel |
WO2022041403A1 (en) * | 2020-08-26 | 2022-03-03 | 中山世达模型制造有限公司 | Sales order prediction method |
CN114708042A (en) * | 2022-05-18 | 2022-07-05 | 北京正汇才科技有限公司 | Financial product scene staging service promotion system and method |
CN114862506A (en) * | 2022-04-24 | 2022-08-05 | 申万宏源证券有限公司 | Financial product recommendation method based on deep reinforcement learning |
WO2022246923A1 (en) * | 2021-05-28 | 2022-12-01 | 中山世达模型制造有限公司 | Method for screening potential customer |
-
2019
- 2019-01-03 CN CN201910005765.7A patent/CN109509040A/en active Pending
Cited By (28)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110147813A (en) * | 2019-04-04 | 2019-08-20 | 深圳价值在线信息科技股份有限公司 | A kind of user draws a portrait construction method, device, storage medium and server |
CN110147813B (en) * | 2019-04-04 | 2021-06-15 | 深圳价值在线信息科技股份有限公司 | User portrait construction method and device, storage medium and server |
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 |
CN110175873A (en) * | 2019-05-24 | 2019-08-27 | 深圳市无忧宝科技有限公司 | Precision Marketing Method and device |
CN110210913A (en) * | 2019-06-14 | 2019-09-06 | 重庆邮电大学 | A kind of businessman frequent customer's prediction technique based on big data |
CN110223166A (en) * | 2019-06-14 | 2019-09-10 | 哈尔滨哈银消费金融有限责任公司 | The prediction technique and equipment of consumer finance user's overdue loan based on big data |
CN110458592A (en) * | 2019-06-18 | 2019-11-15 | 北京海致星图科技有限公司 | Knowledge based map and machine learning algorithm excavate the potential credit client method of bank |
CN110458592B (en) * | 2019-06-18 | 2023-04-07 | 北京海致星图科技有限公司 | Method for mining potential credit clients of bank based on knowledge graph and machine learning algorithm |
CN112308351A (en) * | 2019-07-26 | 2021-02-02 | 上海漕泾热电有限责任公司 | Heat supply balance prediction system and method for cogeneration power plant |
CN112308351B (en) * | 2019-07-26 | 2022-09-09 | 上海漕泾热电有限责任公司 | Heat supply balance prediction system and method for cogeneration power plant |
CN110427555A (en) * | 2019-07-27 | 2019-11-08 | 北京智数时空科技有限公司 | A kind of target user's needing forecasting method and system based on user behavior preference |
CN110688553A (en) * | 2019-08-13 | 2020-01-14 | 平安科技(深圳)有限公司 | Information pushing method and device based on data analysis, computer equipment and storage medium |
CN110674100A (en) * | 2019-09-28 | 2020-01-10 | 上海浦东发展银行股份有限公司信用卡中心 | User demand prediction method and framework based on full-channel operation data |
CN111428092A (en) * | 2020-03-20 | 2020-07-17 | 北京中亦安图科技股份有限公司 | Accurate bank marketing method based on graph model |
CN111783818A (en) * | 2020-04-10 | 2020-10-16 | 南通大学 | Accurate marketing method based on Xgboost and DBSCAN |
CN111783818B (en) * | 2020-04-10 | 2023-09-26 | 南通大学 | Xgboost and DBSCAN-based accurate marketing method |
CN111598580A (en) * | 2020-04-26 | 2020-08-28 | 杭州云象网络技术有限公司 | XGboost algorithm-based block chain product detection method, system and device |
WO2022041403A1 (en) * | 2020-08-26 | 2022-03-03 | 中山世达模型制造有限公司 | Sales order prediction method |
CN112184292A (en) * | 2020-09-16 | 2021-01-05 | 中国农业银行股份有限公司河北省分行 | Marketing method and device based on artificial intelligence decision tree |
CN112256964A (en) * | 2020-10-22 | 2021-01-22 | 重庆邮电大学 | Financial institution potential customer recommendation method based on multi-dimensional data learning |
CN112329928A (en) * | 2020-12-30 | 2021-02-05 | 四川新网银行股份有限公司 | Heterogeneous model-based user satisfaction analysis method |
CN112329928B (en) * | 2020-12-30 | 2021-04-30 | 四川新网银行股份有限公司 | Heterogeneous model-based user satisfaction analysis method |
CN113313572A (en) * | 2021-05-28 | 2021-08-27 | 上海浦东发展银行股份有限公司 | Model identification method based on accumulation fund point-credit customer |
WO2022246923A1 (en) * | 2021-05-28 | 2022-12-01 | 中山世达模型制造有限公司 | Method for screening potential customer |
CN113421133A (en) * | 2021-07-21 | 2021-09-21 | 赛诺数据科技(南京)有限公司 | Network marketing system based on customer matching |
CN114004651A (en) * | 2021-11-02 | 2022-02-01 | 武汉市伟荣发信息科技有限公司 | Information statistics and arrangement system for customer-obtaining channel |
CN114862506A (en) * | 2022-04-24 | 2022-08-05 | 申万宏源证券有限公司 | Financial product recommendation method based on deep reinforcement learning |
CN114708042A (en) * | 2022-05-18 | 2022-07-05 | 北京正汇才科技有限公司 | Financial product scene staging service promotion system and method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109509040A (en) | Predict modeling method, marketing method and the device of fund potential customers | |
CN110796470B (en) | Data analysis system for market subject supervision and service | |
Glaser et al. | Beyond cryptocurrencies-a taxonomy of decentralized consensus systems | |
CN108280541A (en) | Customer service strategies formulating method, device based on random forest and decision tree | |
Chu et al. | Big data analytics for business intelligence in accounting and audit | |
Zhang et al. | Toward a new business reporting model | |
KR102031312B1 (en) | Method for providing p2p fiancial platform based real estate loan service | |
Xue et al. | Cloud computing with AI for banking and e-commerce applications | |
KR102381179B1 (en) | Investment attraction flatform system based on artificial intelligence | |
JP6411562B2 (en) | Information processing apparatus, information processing method, and program | |
Yang et al. | Resilient supply chains to improve the integrity of accounting data in financial institutions worldwide using blockchain technology | |
US20210295436A1 (en) | Method and platform for analyzing and processing investment data | |
Han et al. | ROBO-ADVISORS: REVOLUTIONIZING WEALTH MANAGEMENT THROUGH THE INTEGRATION OF BIG DATA AND ARTIFICIAL INTELLIGENCE IN ALGORITHMIC TRADING STRATEGIES | |
JP2018160258A (en) | Information processor, method for processing information, and program | |
Chan et al. | Artificial intelligence in accounting and auditing | |
Saphyra et al. | A Systematic Literature Review of Fintech Investment and Relationship with Bank in Developed Countries | |
CN115619571A (en) | Financing planning method, system and device | |
Sun et al. | Credit Risk Simulation of Enterprise Financial Management Based on Machine Learning Algorithm | |
Nicoletti | Digital transformation via open data in insurance | |
Veres et al. | The Concept of Using Artificial Intelligence Methods in Debt Financing of Business Entities. | |
Tien | The development of digital banking: a case study of Vietnam | |
Bihari et al. | CRM in Banking Sector with special reference to New Age Banks | |
Oke et al. | The Role of AI in Financial Services: A Bibliometric Analysis | |
Wang | Embedded task system and Gaussian mixture model in the analysis and application of user behavior in marketing management | |
Panja et al. | A Review Report on the Evolution and Implementation of Business Intelligence Technique in the Banking Sector |
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: 20190322 |