CN104751234B - A kind of prediction technique and device of user's assets - Google Patents
A kind of prediction technique and device of user's assets Download PDFInfo
- Publication number
- CN104751234B CN104751234B CN201310753523.9A CN201310753523A CN104751234B CN 104751234 B CN104751234 B CN 104751234B CN 201310753523 A CN201310753523 A CN 201310753523A CN 104751234 B CN104751234 B CN 104751234B
- Authority
- CN
- China
- Prior art keywords
- user
- field
- assets
- prediction model
- field 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.)
- Active
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of prediction techniques of user's assets comprising:Obtain the union feature of the first field user data and the second field user data;First field is financial field, and the second field is not belonging to financial field;According to the asset data and union feature of the first field user, training obtains the assets prediction model of the first field user;The assets prediction model of the first field user is adjusted to the assets prediction model of the second field user using the asset data of the cross-cutting user of predetermined quantity by transfer learning method;According to the assets prediction model of the second field user, the assets of the second field user are predicted.The accuracy that the user's assets prediction model obtained using the method for the present invention is not only predicted is high, but also can reduce cost of labor, improves forecasting efficiency, moreover it is possible to realize lifelong machine learning.
Description
Technical field
The present invention relates to field of information processing, and in particular to a kind of prediction technique and device of user's assets.
Background technology
In fields such as e-commerce, retail, social media and mobile communications, the consumption that service provider has recorded user is inclined
Good, address, action trail, age-sex, social relationships etc., but the asset data of these few field users, it is difficult to seek
Find target user.And in financial field, bank generally possesses user's asset level data and product purchase data (are such as deposited
Money, finance product), address, age-sex, transaction data (such as consumption and payment), withdrawal position data, can using these data
To train the asset level prediction model of bank-user.But this model is difficult to be directly used in the other field users of prediction
The assets of (such as e-commerce user, retailer user, mobile phone user etc.).Because, on the one hand, other field user data and silver
Row user data is not exactly the same, such as bank has user to deposit data, and the enterprise in other fields is without these data;Some
On brand is only wired, under some brands are only wired, the consumption preferences of user can not compare;On the other hand, even if other fields are used
User data and bank-user data feature having the same, there is also assessment differences.For example, the consumption of brand product, uses credit
Block the user of the user and consumption online of payment, asset level may be variant;The user for purchasing by group certain brand on the net may be general
It is low all over the user's asset level for consuming same brand than solid shop/brick and mortar store.
Currently, non-financial business is mainly the asset data by manually marking out non-financial field user, directly training
Prediction model carries out assets prediction to non-financial field user.However, due to the user manually marked asset data amount extremely
It is limited, and accuracy is not high, causes the accuracy that prediction model is predicted relatively low.
Invention content
A kind of prediction technique and device of user's assets are put forward in the embodiment of the present invention, the method for the present invention passes through transfer learning
The assets prediction model of first field user is adjusted to by method using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of second field user.The asset data of the cross-cutting user of the predetermined quantity can be a small amount of number
According to, but the assets prediction model of the second field user is obtained by transfer learning method by the assets prediction model of the first field user
, compared with the assets prediction model generated according to a small amount of user's asset data manually marked, forecasting accuracy
It is improved.
First aspect present invention provides a kind of prediction technique of user's assets, including:
Obtain the union feature of the first field user data and the second field user data;Wherein, the union feature packet
Include the cross feature and cluster feature of the first field user data and the second field user data;First field
For financial field, second field is not belonging to financial field;
According to the asset data of first field user and the union feature, training obtains the money of the first field user
Produce prediction model;
By transfer learning method, using the asset data of the cross-cutting user of predetermined quantity, by first field user
Assets prediction model be adjusted to the assets prediction model of the second field user;Wherein, the cross-cutting user is described first
The co-user of field user and second field user;
According to the assets prediction model of second field user, the assets of second field user are predicted.
In conjunction with first aspect present invention, in the first possible realization method of first aspect, the first field of the acquisition
The union feature of user data and the second field user data, including:
The first field user and second are selected from the first field user data and the second field user data
The cross feature that field user all has;
The feature of the feature of the first field user data and the second field user data is clustered, is obtained
The cluster feature.
In conjunction with first aspect present invention, in second of possible realization method of first aspect, described according to the second neck
The assets prediction model of domain user, after the assets for predicting second field user, the method further includes:
Feedback data is obtained from the application of the assets of second field user;
According to the feedback data, optimize the assets prediction model of second field user.
In conjunction with first aspect present invention, the first or second kind of first aspect may realization method, the of first aspect
In three kinds of possible realization methods, the asset data according to first field user and the union feature, training obtain
The assets prediction model of first field user, including:
The asset data and the union feature of first field user are sampled, the first predetermined quantity is obtained
Sample;
According to the sample of first predetermined quantity, train the assets for the first field user for obtaining the second predetermined quantity pre-
Survey model;
It is described by transfer learning method, using the asset data of the cross-cutting user of predetermined quantity, by first field
The assets prediction model of user is adjusted to the assets prediction model of the second field user, including:
By transfer learning method third predetermined quantity is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user;Wherein, the third predetermined quantity is less than or equal to the second predetermined quantity, integrated
Assets prediction model is the assets prediction model of the second field user.
In conjunction with the third mode in the cards of first aspect present invention, the 4th kind in first aspect is in the cards
It is described that third predetermined quantity is integrated using the asset data of the cross-cutting user of predetermined quantity by transfer learning method in mode
The first field user assets prediction model, including:
By transfer learning method third predetermined quantity is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user, wherein the forecasting accuracy of integrated assets prediction model is higher than by described predetermined
The forecasting accuracy of the assets prediction model for the second field user that the asset data of the cross-cutting user of quantity directly generates.
Second aspect of the present invention provides a kind of prediction meanss of user's assets, including:
Acquisition module, the union feature for obtaining the first field user data and the second field user data;Wherein, institute
State the cross feature and cluster feature that union feature includes the first field user data and the second field user data;
First field is financial field, and second field is not belonging to financial field;
First training module, for according to first field user asset data and the union feature, it is trained
To the assets prediction model of the first field user;
Second training module, will using the asset data of the cross-cutting user of predetermined quantity for passing through transfer learning method
The assets prediction model of first field user is adjusted to the assets prediction model of the second field user;Wherein, described across neck
Domain user is the co-user of first field user and second field user;
Prediction module predicts second field user for the assets prediction model according to second field user
Assets.
In conjunction with second aspect of the present invention, in the first possible realization method of second aspect, the acquisition module is used for
The union feature of the first field user data and the second field user data is obtained, including:
For:Selected from the first field user data and the second field user data the first field user and
The cross feature that second field user all has;And the feature of the first field user data and described second are led
The feature of domain user data is clustered, and the cluster feature is obtained.
In conjunction with second aspect of the present invention, in second of possible realization method of second aspect, described device further includes:
Forecast test module, for, according to the assets prediction model of the second field user, predicting institute in the prediction module
After the assets for stating the second field user, feedback data is obtained from the application of the assets of second field user;According to institute
Feedback data is stated, the assets prediction model of second field user is optimized.
In conjunction with second aspect of the present invention, the first or second kind of second aspect may realization method, the of second aspect
In three kinds of possible realization methods, the first training module is used for the asset data according to first field user and the joint
Feature, training obtain the assets prediction model of the first field user, including:
For:The asset data and the union feature of first field user are sampled, it is predetermined to obtain first
The sample of quantity;According to the sample of first predetermined quantity, training obtains the money of the first field user of the second predetermined quantity
Produce prediction model;
Second training module, will using the asset data of the cross-cutting user of predetermined quantity for passing through transfer learning method
The assets prediction model of first field user is adjusted to the assets prediction model of the second field user, including:
For:By transfer learning method third predetermined number is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user of amount;Wherein, the third predetermined quantity is less than or equal to the second predetermined quantity, collection
At assets prediction model be the second field user assets prediction model.
In conjunction with the third mode in the cards of second aspect of the present invention, the 4th kind in second aspect is in the cards
In mode, the second training module, for by transfer learning method, utilizing the asset data of the cross-cutting user of predetermined quantity, collection
At the assets prediction model of the first field user of third predetermined quantity, including:
For:By transfer learning method third predetermined number is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user of amount, wherein the forecasting accuracy of integrated assets prediction model is higher than by described
The prediction of the assets prediction model for the second field user that the asset data of the cross-cutting user of predetermined quantity directly generates is accurate
Property.
The method of the present invention is by transfer learning method, using the asset data of the cross-cutting user of predetermined quantity, by described
The assets prediction model of one field user is adjusted to the assets prediction model of the second field user.The predetermined quantity it is cross-cutting
The asset data of user can be a small amount of data, without manually marking the asset data of a large amount of user.Therefore, using this hair
The accuracy that user's assets prediction model that bright method obtains not only is predicted is high, but also reduces cost of labor, also improves effect
Rate.
Description of the drawings
Fig. 1 is a kind of prediction technique flow diagram of user's assets;
Fig. 2 is the prediction technique flow diagram of another user's assets;
Fig. 3 is a kind of prediction meanss flow diagram of user's assets;
Fig. 4 is the prediction meanss flow diagram of another user's assets;
Fig. 5 is the prediction meanss flow diagram of another user's assets.
Specific implementation mode
As shown in Figure 1, the embodiment of the present invention provides a kind of prediction technique of user's assets, including:
101, the union feature of the first field user data and the second field user data is obtained;Wherein, the joint is special
Sign includes the cross feature and cluster feature of the first field user data and the second field user data;Described first
Field is financial field, and second field is not belonging to financial field.
In a step 101, first field can be the bank field, second field may include e-commerce,
The fields such as retail, social media and mobile communication.In the joint for obtaining the first field user data and the second field user data
Before feature, the first field user data and the second field user data can be first obtained.Then, from described first
The institute that the first field user and the second field user all have is selected in field user data and the second field user data
State cross feature;The feature of the feature of the first field user data and the second field user data is clustered,
Obtain the cluster feature.
For example, establishing unified feature set according to the user data in bank-user data and other fields.It saves it in
In federated user feature database.The cross feature that user all has in two kinds of data of selection, such as existing credit card purchase also have on the net
The brand of consumption, existing point-of-sale terminal (English:Point of Sale, abbreviation:POS) machine or bank outlets, and have running fix
The place of information.By clustering algorithm, such as the cluster or topic model (English that kmeans algorithms are representative:topic model)
For the Subject Clustering of representative, the various features of two class users are clustered, various features include intersection and Uncrossed feature,
To obtain cluster feature.The classification or topic weights vector of every class sample, the feature as sample can be calculated among these.Such as
Shown in table 1, table 1 shows the feature of bank-user and non-banking user.Wherein, Ub indicates that bank-user, Um indicate non-banking
User, " * " indicate that the data of record, "-" indicate no data.As it can be seen from table 1 bank-user and non-banking user have
About " brand consumption or concern ", the data of " frequent activity venue " and " essential characteristic ", therefore, this three item data belongs to bank
The staggered feature of user data and non-banking user data, " cluster feature weight " have recorded bank-user and non-banking user
Cluster feature.
Table 1
102, according to the asset data of first field user and the union feature, training obtains the first field user
Assets prediction model.
103, by transfer learning method, using the asset data of the cross-cutting user of predetermined quantity, by first field
The assets prediction model of user is adjusted to the assets prediction model of the second field user;Wherein, the cross-cutting user is described
The co-user of first field user and second field user.
For example, TradaBoost algorithms are based on, using the asset data of the cross-cutting user of predetermined quantity, to described first
The assets prediction model of field user carries out transfer learning, and the assets prediction model of first field user is adjusted to second
The assets prediction model of field user.The assets prediction model of the second field user obtained after adjustment with according to a small amount of artificial
The assets prediction model that user's asset data of mark generates is compared, and forecasting accuracy is improved.
Below using the first field user as bank-user, the second field user be consumption on network user for, to the present invention
It illustrates.
As shown in Fig. 2, the embodiment of the present invention also provides a kind of prediction technique of user's assets, including:
201, the union feature of bank-user data and consumption on network user data is obtained;Wherein, the union feature packet
Include the cross feature and cluster feature of the bank-user data and the consumption on network user data.
202, the asset data to the bank-user and the union feature sample, and obtain the first predetermined quantity
Sample.
203, the assets for the bank-user for obtaining the second predetermined quantity according to the sample of first predetermined quantity, training are pre-
Survey model.
Specifically, according to the sample of first predetermined quantity, the multiple graders of training.For example, step 203 includes:It will
User is divided into high assets client and ordinary user as two classes, multiple two graders of training.
204, third predetermined number is integrated using the asset data of the cross-cutting user of predetermined quantity by transfer learning method
The assets prediction model of the bank-user of amount;Wherein, the third predetermined quantity is less than or equal to the second predetermined quantity, integrated
Assets prediction model is the assets prediction model of consumption on network user.
Specifically, increasing the sample of fault discrimination by the asset data of cross-cutting user based on TradaBoost algorithms
Weight obtain integrated classifier to obtain new Weak Classifier, and after taking turns iteration by N.N can be time of setting
Number.Therefore, the embodiment of the present invention can realize the migration of the assets prediction model of bank-user based on TradaBoost algorithms
It practises.
For example, user is divided into high assets client and ordinary user as two classes in step 203, training obtains multiple two points
After class device, TradaBoost algorithms are based on, using the asset data of the cross-cutting user of predetermined quantity, to the first field user
Assets prediction model carries out transfer learning, obtains the integrated classifier for distinguishing bank high assets user and ordinary user.
In step 204, the forecasting accuracy of integrated assets prediction model is higher than by the cross-cutting of the predetermined quantity
The forecasting accuracy of the assets prediction model for the consumption on network user that the asset data of user directly generates.
205, feedback data is obtained from the application of the assets of the consumption on network user.
According to the assets of the consumption on network user of prediction, marketing activity is carried out, and obtain feedback data.The feedback coefficient
According to the asset data that can be consumption on network user.
206, according to the feedback data, optimize the assets prediction model of consumption on network user.
The error of the asset data and feedback data of prediction is calculated, the prediction result of assets prediction model is evaluated and tested.If pre-
It surveys result to report to the leadship after accomplishing a task, then should optimize the assets prediction model of consumption on network user according to the feedback data, realize lifelong engineering
It practises.
For example, bank carries out the user in other fields of the high assets predicted the marketing of high assets threshold finance product
Successful user's mark of marketing is new high assets user by activity.These user data marked can be used for next round
The training of TradaBoost, to optimize the assets prediction model of other field users.
As shown in figure 3, the embodiment of the present invention also provides a kind of prediction meanss 301 of user's assets, including:
Acquisition module 302, the union feature for obtaining the first field user data and the second field user data;Its
In, the union feature includes the cross feature and cluster of the first field user data and the second field user data
Feature;First field is financial field, and second field is not belonging to financial field;
First training module 303 is used for the asset data according to first field user and the union feature, training
Obtain the assets prediction model of the first field user;
Second training module 304, for by transfer learning method, utilizing the assets number of the cross-cutting user of predetermined quantity
According to the assets prediction model of first field user to be adjusted to the assets prediction model of the second field user;Wherein, described
Cross-cutting user is the co-user of first field user and second field user;
Prediction module 305 predicts that second field is used for the assets prediction model according to second field user
The assets at family.
Preferably, the acquisition module 302, the connection for obtaining the first field user data and the second field user data
Feature is closed, including:
For:Selected from the first field user data and the second field user data the first field user and
The cross feature that second field user all has;And the feature of the first field user data and described second are led
The feature of domain user data is clustered, and the cluster feature is obtained.
Preferably, as shown in figure 4, described device 301 further includes:The forecast test module 306, in the basis
The assets prediction model of second field user, after the assets for predicting second field user, from second field user
Assets application in obtain feedback data;According to the feedback data, the assets for optimizing second field user predict mould
Type.
Preferably, the first training module 303, for special according to the asset data of first field user and the joint
Sign, training obtain the assets prediction model of the first field user, including:
For:The asset data and the union feature of first field user are sampled, it is predetermined to obtain first
The sample of quantity;According to the sample of first predetermined quantity, training obtains the money of the first field user of the second predetermined quantity
Produce prediction model;
Second training module 304, for by transfer learning method, utilizing the assets number of the cross-cutting user of predetermined quantity
According to, the assets prediction model of first field user is adjusted to the assets prediction model of the second field user, including:
For:By transfer learning method third predetermined number is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user of amount;Wherein, the third predetermined quantity is less than or equal to the second predetermined quantity, collection
At assets prediction model be the second field user assets prediction model.
Preferably, the second training module 304 utilizes the cross-cutting user of predetermined quantity for by transfer learning method
Asset data integrates the assets prediction model of the first field user of third predetermined quantity, including:
For:By transfer learning method third predetermined number is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user of amount, wherein the forecasting accuracy of integrated assets prediction model is higher than by described
The prediction of the assets prediction model for the second field user that the asset data of the cross-cutting user of predetermined quantity directly generates is accurate
Property.
As shown in figure 5, the embodiment of the present invention also provides a kind of prediction meanss 401 of user's assets, including:Input interface
402, processor 403 and output interface 404, the processor are separately connected input interface 402 and output interface 404.
The processor 403 is used to obtain the union feature of the first field user data and the second field user data;Its
In, the union feature includes the cross feature and cluster of the first field user data and the second field user data
Feature;First field is financial field, and second field is not belonging to financial field;
According to the asset data of first field user and the union feature, training obtains the money of the first field user
Produce prediction model;
By transfer learning method, using the asset data of the cross-cutting user of predetermined quantity, by first field user
Assets prediction model be adjusted to the assets prediction model of the second field user;Wherein, the cross-cutting user is described first
The co-user of field user and second field user;
According to the assets prediction model of second field user, the assets of second field user are predicted.
The processor 403 is used to obtain the union feature of the first field user data and the second field user data, packet
It includes:
For:Selected from the first field user data and the second field user data the first field user and
The cross feature that second field user all has;
The feature of the feature of the first field user data and the second field user data is clustered, is obtained
The cluster feature.
The processor 403 is additionally operable in the assets prediction model according to the second field user, prediction described second
After the assets of field user, feedback coefficient is obtained from the application of the assets of second field user by input interface 402
According to;And for according to the feedback data, optimizing the assets prediction model of second field user.
The processor 403 for according to first field user asset data and the union feature, it is trained
To the assets prediction model of the first field user, including:
For:The asset data and the union feature of first field user are sampled, it is predetermined to obtain first
The sample of quantity;
According to the sample of first predetermined quantity, train the assets for the first field user for obtaining the second predetermined quantity pre-
Survey model;
The processor 403 is used for through transfer learning method, will using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user is adjusted to the assets prediction model of the second field user, including:
For:By transfer learning method third predetermined number is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user of amount;Wherein, the third predetermined quantity is less than or equal to the second predetermined quantity, collection
At assets prediction model be the second field user assets prediction model.
The processor 403 is used to, by transfer learning method, utilize the asset data of the cross-cutting user of predetermined quantity, collection
At the assets prediction model of the first field user of third predetermined quantity, including:
By transfer learning method third predetermined quantity is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user, wherein the forecasting accuracy of integrated assets prediction model is higher than by described predetermined
The forecasting accuracy of the assets prediction model for the second field user that the asset data of the cross-cutting user of quantity directly generates.
One of ordinary skill in the art will appreciate that all or part of step in the various methods of above-described embodiment is can
It is completed with instructing relevant hardware by program, which can be stored in a computer readable storage medium, storage
Medium may include:Read-only memory (ROM, Read Only Memory), random access memory (RAM, Random
Access Memory), disk or CD etc..
The prediction technique and device for being provided for the embodiments of the invention a kind of user's assets above are described in detail,
Principle and implementation of the present invention are described for specific case used herein, and the explanation of above example is only used
In facilitating the understanding of the method and its core concept of the invention;Meanwhile for those skilled in the art, think of according to the present invention
Think, there will be changes in the specific implementation manner and application range, in conclusion the content of the present specification should not be construed as pair
The limitation of the present invention.
Claims (6)
1. a kind of prediction technique of user's assets, which is characterized in that including:
The first field user is selected from the first field user data and the second field user data and the second field user all has
Some cross features, first field are financial field, and second field is not belonging to financial field;
The feature of the feature of the first field user data and the second field user data is clustered, is clustered
Feature;
The asset data and union feature of first field user are sampled, the sample of the first predetermined quantity is obtained,
In, the union feature includes the cross feature and the cluster feature;
According to the sample of first predetermined quantity, training obtains the assets prediction mould of the first field user of the second predetermined quantity
Type;
By transfer learning method the first of third predetermined quantity is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of field user;Wherein, the third predetermined quantity is less than or equal to the second predetermined quantity, by integrated money
Production prediction model is adjusted to the assets prediction model of the second field user;Wherein, the cross-cutting user is first field
The co-user of user and second field user;
According to the assets prediction model of second field user, the assets of second field user are predicted.
2. the prediction technique of user's assets according to claim 1, which is characterized in that described according to the second field user
Assets prediction model, after the assets for predicting second field user, the method further includes:
Feedback data is obtained from the application of the assets of second field user;
According to the feedback data, optimize the assets prediction model of second field user.
3. the prediction technique of user's assets according to claim 1, which is characterized in that described to pass through transfer learning method, profit
With the asset data of the cross-cutting user of predetermined quantity, the assets for integrating the first field user of third predetermined quantity predict mould
Type, including:
By transfer learning method the first of third predetermined quantity is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of field user, wherein the forecasting accuracy of integrated assets prediction model is higher than by the predetermined quantity
Cross-cutting user the forecasting accuracy of the assets prediction model of the second field user that directly generates of asset data.
4. a kind of prediction meanss of user's assets, which is characterized in that including:
Acquisition module, for selecting the first field user and second from the first field user data and the second field user data
The cross feature that field user all has;And by the feature of the first field user data and the second field user number
According to feature clustered, obtain cluster feature, first field be financial field, second field is not belonging to financial neck
Domain;
First training module, for first field user asset data and union feature sample, obtain first
The sample of predetermined quantity;According to the sample of first predetermined quantity, training obtains the first field user of the second predetermined quantity
Assets prediction model, wherein the union feature includes the cross feature and the cluster feature;
Second training module using the asset data of the cross-cutting user of predetermined quantity, integrates for by transfer learning method
The assets prediction model of first field user of three predetermined quantities;Wherein, it is pre- to be less than or equal to second for the third predetermined quantity
Integrated assets prediction model is adjusted to the assets prediction model of the second field user by fixed number amount;Wherein, the cross-cutting use
Family is the co-user of first field user and second field user;
Prediction module predicts the money of second field user for the assets prediction model according to second field user
Production.
5. the prediction meanss of user's assets according to claim 4, which is characterized in that further include:
Forecast test module, in the prediction module according to the assets prediction model of the second field user, prediction described the
After the assets of two field users, feedback data is obtained from the application of the assets of second field user;According to described anti-
Data are presented, the assets prediction model of second field user is optimized.
6. the prediction meanss of user's assets according to claim 4, which is characterized in that the second training module, for passing through
Transfer learning method integrates the first field user of third predetermined quantity using the asset data of the cross-cutting user of predetermined quantity
Assets prediction model, including:
For:By transfer learning method third predetermined quantity is integrated using the asset data of the cross-cutting user of predetermined quantity
The assets prediction model of first field user, wherein the forecasting accuracy of integrated assets prediction model is higher than by described predetermined
The forecasting accuracy of the assets prediction model for the second field user that the asset data of the cross-cutting user of quantity directly generates.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310753523.9A CN104751234B (en) | 2013-12-31 | 2013-12-31 | A kind of prediction technique and device of user's assets |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310753523.9A CN104751234B (en) | 2013-12-31 | 2013-12-31 | A kind of prediction technique and device of user's assets |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104751234A CN104751234A (en) | 2015-07-01 |
CN104751234B true CN104751234B (en) | 2018-10-19 |
Family
ID=53590879
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310753523.9A Active CN104751234B (en) | 2013-12-31 | 2013-12-31 | A kind of prediction technique and device of user's assets |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104751234B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105426534A (en) * | 2015-12-21 | 2016-03-23 | 华为技术有限公司 | Information determination method and device |
CN107066518B (en) * | 2017-02-08 | 2021-01-08 | 创新先进技术有限公司 | Data processing method and system |
CN108734327A (en) * | 2017-04-20 | 2018-11-02 | 腾讯科技(深圳)有限公司 | A kind of data processing method, device and server |
CN113570162A (en) * | 2021-08-30 | 2021-10-29 | 平安医疗健康管理股份有限公司 | Hospitalization cost prediction method and device based on artificial intelligence and computer equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102508859A (en) * | 2011-09-29 | 2012-06-20 | 北京亿赞普网络技术有限公司 | Advertisement classification method and device based on webpage characteristic |
CN102521656A (en) * | 2011-12-29 | 2012-06-27 | 北京工商大学 | Integrated transfer learning method for classification of unbalance samples |
CN102880688A (en) * | 2012-09-14 | 2013-01-16 | 北京百度网讯科技有限公司 | Method, device and equipment for evaluating webpage |
-
2013
- 2013-12-31 CN CN201310753523.9A patent/CN104751234B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102508859A (en) * | 2011-09-29 | 2012-06-20 | 北京亿赞普网络技术有限公司 | Advertisement classification method and device based on webpage characteristic |
CN102521656A (en) * | 2011-12-29 | 2012-06-27 | 北京工商大学 | Integrated transfer learning method for classification of unbalance samples |
CN102880688A (en) * | 2012-09-14 | 2013-01-16 | 北京百度网讯科技有限公司 | Method, device and equipment for evaluating webpage |
Also Published As
Publication number | Publication date |
---|---|
CN104751234A (en) | 2015-07-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104915879B (en) | The method and device that social relationships based on finance data are excavated | |
US10949825B1 (en) | Adaptive merchant classification | |
JP7153722B2 (en) | Automated enterprise transaction data aggregation and accounting | |
CN107578281A (en) | User preferential certificate behavior prediction method and model building method under e-commerce environment | |
CN105490823B (en) | data processing method and device | |
Molloy et al. | Graph analytics for real-time scoring of cross-channel transactional fraud | |
KR102297669B1 (en) | System for providing matching service for connecting between manufacturer and distributor | |
Duman et al. | A novel and successful credit card fraud detection system implemented in a turkish bank | |
CN108628866B (en) | Data fusion method and device | |
CN111325619A (en) | Credit card fraud detection model updating method and device based on joint learning | |
US20140229233A1 (en) | Consumer spending forecast system and method | |
CN109598542A (en) | A kind of put-on method, device and the electronic equipment of equity of marketing | |
CN110033120A (en) | For providing the method and device that risk profile energizes service for trade company | |
CN104751234B (en) | A kind of prediction technique and device of user's assets | |
US20160071133A1 (en) | Automatic determination of account owners to be encouraged to utilize point of sale transactions | |
CN108416684A (en) | A kind of credibility appraisal procedure, device and the server of account main body | |
US20160005060A1 (en) | Method and system for predicting spending on travel | |
Yuping et al. | New methods of customer segmentation and individual credit evaluation based on machine learning | |
CN109919667A (en) | A kind of method and apparatus of the IP of enterprise for identification | |
CN112330373A (en) | User behavior analysis method and device and computer readable storage medium | |
CN109858952A (en) | Data processing method and device under service scenarios | |
CN105512914A (en) | Information processing method and electronic device | |
CN104839962A (en) | Smart wallet, information processing method thereof and device | |
Tang et al. | Research on e-commerce customer churn based on RFM model and naive Bayes algorithm | |
Polkowski et al. | BIG DATA APPLICATIONS IN SMES. |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |