CN109242672A - Refund information forecasting method, device and the computer readable storage medium of loan - Google Patents
Refund information forecasting method, device and the computer readable storage medium of loan Download PDFInfo
- Publication number
- CN109242672A CN109242672A CN201811151423.8A CN201811151423A CN109242672A CN 109242672 A CN109242672 A CN 109242672A CN 201811151423 A CN201811151423 A CN 201811151423A CN 109242672 A CN109242672 A CN 109242672A
- Authority
- CN
- China
- Prior art keywords
- stages
- information
- history
- loan
- refund
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Technology Law (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The embodiment of the present application provide the refund information forecasting method of loan a kind of, device and and computer readable storage medium, this method comprises: according to the first history refund information of the first of loan the first refund information by stages and the first corresponding loan by stages, the second history refund information of the second of the loan the corresponding loan by stages is obtained；This first by stages for this second by stages upper one by stages；The first refund information is according to the portrait information of the user of the loan, the first history refund information and machine learning model trained in advance, and the user predicted is directed to the first refund information by stages；According to the portrait information, the second history refund information and the machine learning model, predict the user for the second second refund information by stages.The application realizes the prediction to the refund information by stages that do not refund respectively of loan.
Description
Technical field
This application involves field of computer technology, in particular to the refund information forecasting method of a kind of loan, device and
Computer readable storage medium.
Background technique
With the development of internet technology, internet financial field is flourished.Various payments, lending platforms and application are not
It is disconnected to emerge in large numbers, it is more and more convenient for people's lives.Network loan is internet monetary items more concerned in recent years.Network is borrowed
Money can more easily provide the fund of urgent need for user, and user is facilitated to live.
Due to credit authorization higher cost, network loan is not stringent generally for the credit audit of user, thus brings height
The problem of overdue rate and fraction defective.Refund prediction is carried out during loan for loan user, is conducive to network loan platform
User is pressed for payment of to refund, avoid risk and reasonable disposition fund etc. plays an important role.
Currently, it is badly in need of a kind of method of refund information by stages that do not refund respectively that can predict loan, it is above-mentioned to realize
Effect.
Summary of the invention
The embodiment of the present application provides refund information forecasting method, device and the computer-readable storage medium of a kind of loan
Matter can predict the refund feelings information by stages that do not refund respectively of loan.
In a first aspect, the embodiment of the present application provides a kind of refund information forecasting method of loan, comprising:
According to the first of loan the first refund information by stages and described first by stages the corresponding loan first go through
History refund information obtains the second history refund information of the second of the loan the corresponding loan by stages；Described first point
Phase be described second by stages upper one by stages；The first refund information is portrait information according to the user of the loan, institute
The first history refund information and machine learning model trained in advance are stated, the user predicted is directed to described first point
The refund information of phase；
According to the portrait information, the second history refund information and the machine learning model, the user is predicted
For described second the second refund information by stages.
In the program, by upper one the first refund information and the first history refund information by stages according to prediction, obtain
Currently the second history refund information by stages, further according to the portrait information of the user of the loan, the second history refund information and pre-
First trained machine learning model, current the second refund information by stages of prediction, not refunding respectively for loan can be predicted by realizing
Refund information by stages purpose.
It is described according to the portrait information, the second history refund information and the machine in a kind of possible design
Device learning model predicts the user for described second the second refund information by stages, comprising:
By machine learning model described in the portrait information, the second history refund information input, a pre- direction finding is exported
Amount, the predicted vector are used to indicate the second refund information.
In a kind of possible design, the component of the predicted vector include the user for described second by stages also
The overdue number of days of money is located at each probability preset in overdue section；
If the maximum probability in each probability is greater than preset threshold, the second refund information includes the user for institute
It states the overdue number of days of the second refund by stages and is located at target and preset in overdue section；Wherein, the target presets overdue section as institute
State that maximum probability is corresponding to preset overdue section.
In a kind of possible design, the first history refund information includes: the described first corresponding loan by stages
First history refund statistical information of money, the second history refund statistical information include the described second corresponding loan by stages
Second history refund statistical information of money；It is described by stages right according to the first of loan the first refund information by stages and described first
First history refund information of the loan answered obtains the second history of the second of the loan the corresponding loan by stages
Refund information, comprising:
Overdue section and the first history refund statistical information are preset according to the corresponding target of the first refund information,
Obtain the second history refund statistical information；
According to the second history refund statistical information, the second history refund information is obtained.
In a kind of possible design, the first history refund information includes: the described first corresponding loan by stages
The refund information of the history of money each phase；The second history refund information includes: the described second corresponding loan by stages
The refund information of history each phase；It is described according to the first of loan the first refund information by stages and the described first corresponding institute by stages
The the first history refund information for stating loan, the second history for obtaining the second of the loan the corresponding loan by stages are refunded letter
Breath, comprising:
Overdue section and the described first corresponding loan by stages are preset according to the corresponding target of the first refund information
The refund information of the history of money each phase obtains the refund information of described second history each phase of the corresponding loan by stages；
According to the refund information of described second history each phase of the corresponding loan by stages, second history is obtained also
Money information.
In a kind of possible design, the first history refund information includes: the described first corresponding loan by stages
The refund information of first history refund statistical information of money and described first history each phase of the corresponding loan by stages, it is described
Second history refund information includes the described second second history refund statistical information of the corresponding loan and described by stages
The refund information of two history each phases of the corresponding loan by stages；The first first refund information by stages according to loan
With the first history refund information of the described first corresponding loan by stages, the second of the loan the corresponding institute by stages is obtained
State the second history refund information of loan, comprising:
Overdue section and the first history refund statistical information are preset according to the corresponding target of the first refund information, is obtained
Second history refund statistical information；
Overdue section and the described first corresponding loan by stages are preset according to the corresponding target of the first refund information
The refund information of the history of money each phase obtains the refund information of described second history each phase of the corresponding loan by stages；
According to the second history refund statistical information and described second history each phase of the corresponding loan by stages
Refund information obtains the second history refund information.
In a kind of possible design, the method also includes:
Multiple training samples are obtained, the training sample includes that once providing a loan for training user is each by stages corresponding
The information of training user, the information of the training user include that the portrait information of the training user is gone through with accordingly by stages corresponding
History refund information；
Obtain the label of each training sample, the label be used to indicate corresponding training sample it is corresponding once provide a loan it is each
Corresponding refund information by stages；
The training of the multiple training sample is obtained according to the label of the multiple training sample and each training sample
The machine learning model.
In a kind of possible design, the training user is matches with the user, and loan types and the loan
The user that the type of money matches.
Second aspect, the embodiment of the present application provide a kind of refund information prediction device of loan, comprising:
Module is obtained, for by stages corresponding described according to the first of loan the first refund information by stages and described first
First history refund information of loan, the second history for obtaining the second of the loan the corresponding loan by stages are refunded letter
Breath；Described first by stages for second by stages upper one by stages；The first refund information is the picture according to the user of the loan
As information, the first history refund information and machine learning model trained in advance, the user predicted are directed to
Described first refund information by stages；
Prediction module is used for according to the portrait information, the second history refund information and the machine learning model,
Predict the user for described second the second refund information by stages.
In a kind of possible design, the prediction module is specifically used for:
By machine learning model described in the portrait information, the second history refund information input, predicted vector is exported,
The predicted vector is used to indicate the second refund information.
In a kind of possible design, the component of the predicted vector include the user for described second by stages also
The overdue number of days of money is located at each probability preset in overdue section；
If the maximum probability in each probability is greater than preset threshold, the second refund information includes the user for institute
It states the overdue number of days of the second refund by stages and is located at target and preset in overdue section；Wherein, the target is preset described in overdue section
Maximum probability is corresponding to preset overdue section.
In a kind of possible design, the first history refund information includes the described first corresponding loan by stages
The first history refund statistical information, the second history refund statistical information includes the described second corresponding loan by stages
The second history refund statistical information；The acquisition module, is specifically used for:
Overdue section and the first history refund statistical information are preset according to the corresponding target of the first refund information,
Obtain the second history refund statistical information；
According to the second history refund statistical information, the second history refund information is obtained.
In a kind of possible design, the first history refund information includes: the described first corresponding loan by stages
The refund information of the history of money each phase；The second history refund information includes: the described second corresponding loan by stages
The refund information of history each phase；The acquisition module, is specifically used for:
Overdue section and the described first corresponding loan by stages are preset according to the corresponding target of the first refund information
The refund information of the history of money each phase obtains the refund information of described second history each phase of the corresponding loan by stages；
According to the refund information of described second history each phase of the corresponding loan by stages, second history is obtained also
Money information.
In a kind of possible design, the first history refund information includes: the described first corresponding loan by stages
The refund information of first history refund statistical information of money and described first history each phase of the corresponding loan by stages, it is described
Second history refund information includes the described second second history refund statistical information of the corresponding loan and described by stages
The refund information of two history each phases of the corresponding loan by stages；The acquisition module, is specifically used for:
Overdue section and the first history refund statistical information are preset according to the corresponding target of the first refund information, is obtained
Second history refund statistical information；
Overdue section and the described first corresponding loan by stages are preset according to the corresponding target of the first refund information
The refund information of the history of money each phase obtains the refund information of described second history each phase of the corresponding loan by stages；
According to the second history refund statistical information and described second history each phase of the corresponding loan by stages
Refund information obtains the second history refund information.
It further include training module in a kind of possible design；The training module is used for:
Multiple training samples are obtained, the training sample includes that once providing a loan for training user is each by stages corresponding
The information of training user, the information of the training user include that the portrait information of the training user is gone through with accordingly by stages corresponding
History refund information；
Obtain the label of each training sample, the label be used to indicate corresponding training sample it is corresponding once provide a loan it is each
Corresponding refund information by stages；
The training of the multiple training sample is obtained according to the label of the multiple training sample and each training sample
The machine learning model.
The third aspect, the embodiment of the present application provide a kind of readable storage medium storing program for executing, including program or instruction, when described program or
When instruction is run on computers, the method in first aspect and any possible design of first aspect is performed.
Fourth aspect, the embodiment of the present application provide a kind of electronic equipment, comprising: processor, the processor and memory
Coupling；
The memory is used for, and stores computer program；
The processor is used for, and calls the computer program stored in the memory, to realize first aspect and
On the one hand the method in any possible design.
By upper one the first refund information and the first history refund information by stages according to prediction in the application, worked as
Preceding the second history refund information by stages, further according to the portrait information of the user of the loan, the second history refund information and in advance
Trained machine learning model, prediction current the second refund information by stages, not refunding respectively for loan can be predicted by realizing
The purpose of refund information by stages.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is this Shen
Some embodiments please for those of ordinary skill in the art without any creative labor, can be with
It obtains other drawings based on these drawings.
Fig. 1 is the flow chart of the refund information forecasting method of loan provided by the embodiments of the present application；
Fig. 2 is the flow chart of the acquisition methods of machine learning model provided by the embodiments of the present application；
Fig. 3 is the structural schematic diagram one of the refund information prediction device of loan provided by the embodiments of the present application；
Fig. 4 is the structural schematic diagram two of the refund information prediction device of loan provided by the embodiments of the present application；
Fig. 5 is the structural schematic diagram of electronic equipment provided by the embodiments of the present application.
Specific embodiment
To keep the purposes, technical schemes and advantages of the embodiment of the present application clearer, below in conjunction with the embodiment of the present application
In attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is
Some embodiments of the present application, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall in the protection scope of this application.
Aiming at the problem that can be carried out prediction currently without the refund situation by stages that do not refund respectively to loan, this is proposed
Scheme.This programme can be not only used for the prediction provided a loan on the net, also can be applied to the prediction provided a loan under line.
Some embodiments of the present application are described below with reference to Fig. 1.Fig. 1 is that the refund of loan provided by the embodiments of the present application is believed
The flow chart of prediction technique is ceased, referring to Fig. 1, the method for the present embodiment includes:
Step S101, according to the first of loan the first refund information by stages and this first by stages the corresponding loan
One history refund information obtains the second history refund information of the second of the loan the corresponding loan by stages；This is first by stages
For this second by stages upper one by stages；The first refund information is portrait information, first history according to the user of the loan
Refund information and machine learning model trained in advance, the user predicted are directed to the first refund information by stages；
Step S102, according to the portrait information, the second history refund information and the machine learning model, the user is predicted
For the second second refund information by stages.
Specifically, the method for the present embodiment can be realized based on the refund information prediction device of loan.Wherein, the refund of loan
Information prediction device can be based on hardware or software realization.Loan involved in this embodiment not only includes traditional on line
Or the loaning bill behavior of Xian Xia credit agency, the behaviors such as monthly payment plan, credit card installment reimbursement when can also include purchase commodity, all
Can whether can be overdue according to the refund of the program prediction user of the application.
In step s101, according to the first of loan the first refund information by stages and the first corresponding loan by stages
The first history refund information, obtain the second history refund information of the second of the loan the corresponding loan by stages.
Specifically, second in the present embodiment by stages can for the loan it is corresponding it is multiple by stages in do not refund in addition to first
By stages other than it is any by stages, first by stages for second by stages upper one by stages.Such as: the loan be divided into 6 by stages into
Row is refunded, wherein it 1 and by stages 2 has refunded by stages, then, if second is by stages by stages 4, first is by stages by stages 3, if
Second is by stages by stages 5, then first is by stages by stages 4.It is understood that if first is by stages by stages 1, first by stages
First history refund information can be full 0 value.
It, can be in conjunction with the portrait information and history refund information one of user if producing history refund information for the loan
Play the prediction for corresponding refund information by stages.History refund information may include: the refund information and history of history each phase
At least one of refund statistical information.
The refund information of history each phase includes: the compensatory amount of money of overdue number of days, lending agency, the loan of history each phase of loan
At least one of compensatory ratio of mechanism.Refund by stages for example, being divided into 6 with the loan, for by stages 3, by stages 1 and point
The respective refund information of phase 2 is the refund information of 3 accordingly history each phases by stages.Wherein, lending agency is compensatory refers to user
Lending agency pays for the behavior of refund first to fund side after overdue certain time.
History refund statistical information includes: the compensatory total gold of the overdue total degree of history, overdue total amount, lending agency of loan
At least one of the compensatory total degree of volume, lending agency, the compensatory ratio of lending agency, maximum overdue number of days, minimum overdue number of days.It borrows
The history refund statistical information of money is that can also be selected according to actual needs the statistical information of the refund information of history each phase
It takes, is not limited to examples cited.Such as: for by stages 3, the respective refund information in by stages 1 and by stages 2 is counted to get arriving
3 corresponding history refund statistical information by stages.
Based on this, the first history refund information of the first corresponding loan by stages may include: first by stages corresponding
At least one of the history refund statistical information of the refund information of the history of the loan each phase and the first corresponding loan by stages.
Second history refund information of the second corresponding loan by stages may include: history each phase of the second corresponding loan by stages
Refund information and the second corresponding loan by stages at least one of history refund statistical information.
Illustratively, if first is by stages by stages the 2 of the loan, the history of the first corresponding loan by stages is refunded letter
Breath includes the statistical information of the 1 refund information and the refund information to by stages 1 by stages of the loan；If first is the loan by stages
By stages 4, then the history refund information of the first corresponding loan by stages includes the by stages 1, by stages 2 and by stages 3 each of the loan
From refund information and the respective refund information to by stages 1, by stages 2 and by stages 3 statistical information.If first is the loan by stages
By stages the 1 of money, then first by stages without history refund information, i.e., the first history for including in the first user information at this time is refunded letter
Breath can be the value of the no history refund information of instruction, such as 0 value.
For refund information: refund information is corresponding refund situation by stages of the user for the loan of the loan.
Possible refund situation are as follows: 0 day overdue (there is no overdue, i.e., normal to refund), overdue number of days is less than or equal to M1, overdue number of days
Greater than M1 and it is less than or equal to M2, overdue number of days is greater than M2 and is less than or equal to M3, and overdue number of days is greater than M3 days or more etc..Wherein,
It can be 60, M3 can be 90 that M1, which can be 30, M2,.
So, first the first refund information by stages, the as user of the loan for the loan first by stages also
Money situation.First the first refund information by stages be according to the portrait information of the user of the loan, this first by stages
One history refund information and machine learning model trained in advance, the user of the loan predicted for this first by stages
Refund information.Wherein, for the portrait information of the user of the loan, comprising: the personal letter such as age, gender of the user of the loan
Breath, bank card are opened an account the time, at least one of personal credit informations such as credit grade.The personal information such as age, gender can be with base
The registration information when platform of the user's registration loan obtains.The credit informations such as the credit grade of bank card these can be from outer
It is obtained in portion's data source, such as the external channels such as credit agency's system or e-commerce platform.Specific user's portrait information
It can be selected according to actual needs.It is understood that for the loan each by stages for, it is respectively by stages corresponding
User's portrait information is identical.
Specifically, which, which can be, offline obtains training sample training.Wherein, the machine learning mould
Type can be Recognition with Recurrent Neural Network (recurrent neural networks, abbreviation RNN) model, alternatively, long short-term memory is neural
Network (long short termmemory, LSTM) model, alternatively, bidirectional circulating neural network (bidirections
Recurrent neural networks, abbreviation BRNN) model.
Obtain the process of first the first refund information by stages are as follows: by the portrait information of the user of the loan, this
First history refund information input machine learning model of the one corresponding loan by stages, exports the first predicted vector, this
One predicted vector is used to indicate the first first refund information by stages.
If the component of first predicted vector includes the user of the loan for the overdue day numerical digit of first refund by stages
In each probability preset in overdue section；And if the maximum probability in each probability is greater than preset threshold, the first refund packet
The user for including the loan is located at the second target for the overdue number of days of first refund by stages and presets in overdue section, wherein should
Second target presets that the maximum probability in overdue the first predicted vector of section is corresponding to preset overdue section.That is the first refund information
It is corresponding that overdue section is preset with the second target.
For example, being provided with M presets overdue section, then the first predicted vector exported can include at least following each component
(P1... ..., Pm..., PM), wherein PmThe user of the loan is used to indicate for the overdue day numerical digit of first refund by stages
In the probability that m-th is preset in overdue section, 1≤m≤M.
Illustratively, it is each preset overdue section be 0, (0,30], (30,60], (60,90], (90 ,+∞), indicate respectively as
Lower refund information: 0 day overdue (normal refund), overdue number of days be section (0,30] in (overdue number of days is less than or equal to 30 within one day
It), overdue number of days be section (30,60] in one day (overdue number of days be greater than 30 days and be less than or equal to 60 days), overdue number of days is
Section (60,90] in one day (overdue number of days be greater than 60 days and be less than or equal to 90 days), overdue number of days be section (90 ,+∞) in
One day (overdue number of days be greater than 90 days).
If in the case where this is exemplary, the first predicted vector of output at least partially (0.05,0.05,0.8,0.04,
0.06), preset threshold 0.7, then the first refund information predicted are as follows: the user of the loan exceedes for first refund by stages
Phase number of days be located at preset overdue section (30,60] in-overdue number of days is greater than 30 days and is less than or equal to 60 days；At this point, presetting overdue
Section (30,60] it is that the second target presets overdue section, i.e. corresponding second target of the first refund information presets overdue section and is
(30,60].
In addition, the first predicted vector may also include this first by stages the compensatory amount of money of corresponding lending agency be located at each default generation
Probability between indemnity frontal region and/or this first by stages the compensatory ratio of corresponding lending agency be located at and each preset compensatory ratio section
In probability.At this point, first the first refund information by stages is in addition to including: that refund overdue number of days is located at maximum probability corresponding the
Two targets are preset in overdue section, further includes: first to be located at the second target default for the compensatory amount of money of corresponding lending agency by stages for this
In commutation frontal region and/or this first by stages the compensatory ratio of corresponding lending agency be located at the second target and preset compensatory ratio area
In.
After having obtained the first refund information, then, according to the first of loan the first refund information by stages and this first
First history refund information of the corresponding loan by stages obtains the second history of the second of the loan the corresponding loan by stages
Refund information.Below to " according to the first of loan the first refund information by stages and this first by stages the corresponding loan the
One history refund information obtains the second history refund information of the second of the loan the corresponding loan by stages " it is illustrated.
In the first way, if first the first history refund information by stages includes: that this is first by stages corresponding
First history refund statistical information of the loan, the second history refund statistical information include the second corresponding loan by stages
The second history refund statistical information；According to first the first history refund information by stages and this first by stages this first
Refund information obtains the second history refund information of the second corresponding loan by stages, comprising:
Overdue section and first history are preset according to first corresponding second target of the first refund information by stages
Refund statistical information obtains the second history refund statistical information；According to the second history refund statistical information, obtains this and second go through
History refund information.
Specifically, in this way, presetting overdue section due to the second target can indicate that the user of the loan is directed to
The first refund information by stages, therefore, can be preset according to corresponding second target of the first refund information overdue section and this
One the first history refund statistical information by stages obtains the second second history refund statistical information by stages, according to this second
The second history refund statistical information by stages, obtains the second history refund information.
Such as: if including that history adds up overdue number of days, history adds up overdue number, history most in history refund statistical information
The minimum overdue number of days of big overdue number of days, history, then it is pre- according to first corresponding second target of the first refund information by stages
If the overdue number of days of overdue section instruction and whether overdue updating first corresponding history adding up overdue number of days by stages, history adds up
The maximum overdue number of days of overdue number, history, the minimum overdue number of days of history, just obtain this second by stages accordingly history add up it is overdue
Number of days, history add up the maximum overdue number of days of overdue number, history, the minimum overdue number of days of history.
Illustratively, if first corresponding second target of the first refund information by stages preset overdue section be (30,
60], then illustrate overdue number of days be greater than 30 days and be less than or equal to 60 days, then should (30,60] instruction overdue number of days can be 31~60
In any one day, such as (30,60] instruction overdue number of days be 31, then second by stages corresponding history add up overdue number of days etc.
In first by stages corresponding history add up overdue number of days plus 31；Should (30,60] instruction whether it is overdue be " overdue ", then second
Corresponding history adds up overdue number corresponding history adds up overdue number and adds 1 by stages equal to first by stages；By stages for second
The maximum overdue number of days of corresponding history, the minimum overdue number of days of history, then according to first by stages before the loan it is each by stages
First overdue number of days 31 by stages of overdue number of days and prediction, therefrom selects maximum value, and as second, corresponding history is most by stages
Big overdue number of days, therefrom selects minimum value, as the minimum overdue number of days of history.
If in history refund statistical information further including the compensatory total amount of lending agency and/or the compensatory toatl proportion of lending agency,
Can then be obtained according to the first predicted vector the second target preset commutation frontal region between and/or the second target preset compensatory ratio area
Between, the second target is preset between commutation frontal region, the second target is preset the acquisition methods in compensatory ratio section and default exceeded referring to second
The acquisition methods in phase section.So, in advance according to the first history refund statistical information corresponding with the first refund information second
If between overdue section and the default commutation frontal region of the second target and/or the second target presets compensatory ratio section, second is obtained
The second history refund statistical information by stages.Wherein, according to first corresponding second target of the first refund information by stages
It is second by stages right to obtain this for the compensatory amount of money indicated between default commutation frontal region and the first total compensatory amount of money of corresponding history by stages
The total compensatory amount of money of the history answered, presets compensatory ratio area according to first corresponding second target of the first refund information by stages
Between the compensatory ratio that indicates and the first total compensatory ratio of corresponding history by stages, obtain this second by stages corresponding history it is always compensatory
Ratio.And the second target presets the compensatory amount of money indicated between commutation frontal region, the second target presets compensatory ratio section instruction
The acquisition methods of compensatory ratio preset the acquisition methods of the overdue number of days of overdue section instruction referring to the second target.
In the second way, if first the first history refund information by stages includes: that this is first by stages corresponding
The refund information of the history of the loan each phase；The second history refund information include: this second by stages the corresponding loan go through
The refund information of history each phase；According to first the first history refund information by stages and first the first refund letter by stages
Breath obtains the second history refund information of the second corresponding loan by stages, comprising:
According to first corresponding second target of the first refund information by stages preset overdue section and this first by stages
The refund information of the history of the corresponding loan each phase, obtain this second by stages history each phase of the corresponding loan refund letter
Breath；According to the refund information of second history each phase of the corresponding loan by stages, the second history refund information is obtained.
Specifically, the refund information of second history each phase of the corresponding loan by stages include: this first by stages before
Corresponding refund information of each phase, and first refund information by stages of prediction, i.e., in first corresponding loan by stages
On the basis of the refund information of the history of money each phase, add the first refund information by stages of prediction to get arrived this second
The refund information of the history of the corresponding loan each phase by stages.
In the third mode, first the first history refund information by stages include: this first by stages it is corresponding should
The refund information of first history refund statistical information of loan and first history each phase of the corresponding loan by stages, this second
History refund information include this second by stages the second history refund statistical information of the corresponding loan and this second correspond to by stages
The loan history each phase refund information；According to first the first history refund information by stages and this first by stages
The first refund information obtains the second history refund information of the second corresponding loan by stages, comprising:
Overdue section and the first history are preset also according to first corresponding second target of the first refund information by stages
Money statistical information obtains the second history refund statistical information；According to first the first refund information by stages corresponding
Two targets preset the refund information in overdue section and first history each phase of the corresponding loan by stages, obtain this second by stages
The refund information of the history of the corresponding loan each phase；According to the second history refund statistical information and this is second by stages corresponding
The refund information of the history of the loan each phase obtains the second history refund information.
The acquisition methods of second the second history refund information by stages in which are referring to saying in above two mode
Bright, details are not described herein again.
In step s 102, according to the portrait information of the user of the loan, the second history refund information and the machine
Learning model predicts the user of the loan for the second second refund information by stages.
Specifically: by the portrait information of the user of the loan and the second history refund information input machine learning mould
Type, exports the second predicted vector, which is used to indicate the second second refund information by stages.
If the component of second predicted vector includes the user of the loan for the overdue day numerical digit of second refund by stages
In each probability preset in overdue section；And if the maximum probability in each probability is greater than preset threshold, the second refund packet
The user for including the loan is located at first object for the overdue number of days of second refund by stages and presets in overdue section, wherein should
First object presets overdue section and presets overdue section for the maximum probability in the second predicted vector is corresponding.
Specific implementation for the step, it is no longer superfluous herein referring to the acquisition process of first the first refund information by stages
It states.
In conclusion by being corresponded to according to first the first refund information and first by stages of prediction by stages in the present embodiment
The loan the first history refund information, obtain the second history refund information of the second of the loan the corresponding loan by stages,
First by stages for second by stages upper one by stages；Portrait information, the second history refund information further according to the user of the loan
Machine learning model trained in advance, the second refund information of prediction second by stages, realize can predict loan it is each not
The purpose for the refund information by stages refunded.
It should be understood that magnitude of the sequence numbers of the above procedures are not meant that the order of the execution order, the execution of each process is suitable
Sequence should be determined by its function and internal logic, and the implementation process without coping with the embodiment of the present application constitutes any restriction.
The acquisition methods of the machine learning model in a upper embodiment are illustrated using specific embodiment below.Figure
2 be the flow chart of the acquisition methods of machine learning model provided by the embodiments of the present application.Referring to fig. 2, the method for the present embodiment, packet
It includes:
Step S201, multiple training samples are obtained, which includes that once providing a loan for training user is each by stages each
The information of self-corresponding training user, the information of training user include that the portrait information of training user is gone through with accordingly by stages corresponding
History refund information；
Step S202, the label of each training sample is obtained, it is corresponding primary which is used to indicate corresponding training sample
Each by stages corresponding refund information of loan；
Step S203, according to the label of multiple training sample and each training sample, multiple training sample is trained,
Obtain machine learning model.
Specifically, the acquisition methods of machine learning model can be realized based on the acquisition device of machine learning model, engineering
The refund information prediction device of the acquisition device and loan of practising model can be located in same equipment, may be alternatively located at different equipment
In.
Training user in the present embodiment can be to match with the user of the loan in embodiment shown in FIG. 1, and provides a loan
The user that the type of loan in type and embodiment shown in FIG. 1 matches；Wherein, the loan in embodiment shown in FIG. 1
User be user to be predicted.
Alternative user and user to be predicted can be matched by portrait information and history credit information.For example, choosing
Take the age gap of age and user to be predicted within a preset range, the region distance of region and user to be predicted is in pre-determined distance model
In enclosing, the rank difference of credit grade and user to be predicted is within the scope of predetermined level, the amount of loan limit and user to be predicted
The alternative user that difference waits within the scope of default amount is as training user, the personal attribute of these alternative users and user to be predicted
Feature is similar, and loan types are similar, based on these users obtained model of training when carrying out the prediction of the overdue behavior of user more
It is accurate to add.Specific matching principle can be configured according to actual needs, for example, predicting for petty load user
When, it can choose that the age is similar, alternative user of the identical petty load of city of residence is as training user.
Training user and training user's relevant information can be obtained from external data source, can also be from internal data source
Middle acquisition, or obtained jointly in conjunction with external data source with internal data source.It is seldom in internal data source, such as at the beginning of loan platform
In the case that phase makes loans, after being trained using external data source to model, it can directly carry out model migration and user is carried out
Predict risk online to fast implement model, and predicting accurate, the initial stage that reduces makes loans.With the continuous increasing of internal data
It is more, it can use internal data correction model.Further, when internal data is enough, it can use and instructed based on internal data
The model got directly predicts the overdue behavior of user.
External data source is for example including at least one of external credit agency's system and e-commerce platform channel.From outside
The history refund information for the user that data source obtains may include that the refund information of history loan, credit card Historical Stages are refunded
Information or the monthly payment plan information etc. for buying commodity, these information may be used to the training of model.
It obtains and is somebody's turn to do according to above-mentioned training user's relevant information in step S201, for the primary loan of a training user
The once information to each by stages respective training user of money, obtains a training sample.Wherein, the information of training user includes
The portrait information of training user and corresponding by stages corresponding history refund information, the item and upper one that portrait information herein includes is in fact
Apply that the item that the portrait information in example includes is identical, the history in the item that history refund information herein includes and a upper embodiment is also
The item that money information includes is identical.
According to above-mentioned identical method, sufficient amount of training sample is obtained, each training sample includes training user
Each by stages corresponding training user once to provide a loan information.It is understood that if a training user carries out
Repeatedly loan, then can be obtained multiple training samples according to the credit information of the training user.
In step S202, the label of each training sample is obtained, it is corresponding which is used to indicate corresponding training sample
Each by stages corresponding refund information once provided a loan；
If it includes by stages 4 vectors in the label of the training sample that the corresponding primary loan of training sample, which includes 4,
The content for each component instruction that the content and the predicted vector in a upper embodiment for each component instruction that each vector includes include
Identical, each component that sequence of each component that each vector includes in vector includes with the predicted vector in a upper embodiment exists
Sequence in vector is identical.
Such as: if the user that the component of the predicted vector of a upper embodiment includes the loan exceedes for corresponding refund by stages
Phase number of days is located at each probability preset in overdue section, and each to preset overdue section be 0, and (0,30], (30,60], (60,90],
(90 ,+∞)；If corresponding 4 once to provide a loan of training sample by stages in by stages 1 overdue number of days be 0, by stages 2 it is overdue
Number of days is 70 days, and 3 overdue number of days is 20 days by stages, and 4 overdue number of days is 0 by stages, then can be in 1 corresponding vector by stages
(1,0,0,0,0) can be (0,0,1,0,0) in 2 corresponding vectors by stages, can be in 3 corresponding vectors by stages (0,1,0,0,
0), can be in 4 corresponding vectors by stages (1,0,0,0,0), then the label of the training sample can include: (1,0,0,0,0), (0,
0,1,0,0), (0,1,0,0,0) and (1,0,0,0,0).
In step S203, according to the label of multiple training sample and each training sample, multiple training sample is instructed
Practice, obtains machine learning model.
If machine learning model is RNN model, according to the label of multiple training samples and each training sample, use
RNN algorithm is trained multiple training sample, can obtain machine learning model-RNN model.RNN algorithm is existing
Algorithm, repeat no more in the present embodiment.
Method through this embodiment, has got machine learning model, for loan respectively do not refund distinguish it is pre-
It surveys.
It is understood that in above-described embodiment, the number of data or training user either for user to be predicted
According to before the use, requiring after being pre-processed just to obtain input machine learning model.For example, preprocessing process is successively
Are as follows: date amendment of refunding, data accuracy verification, statistics, the data normalization etc. of history refund information.Refund date fixed case
It such as include: big numerical quantity missing or the amendment in responsible processing business data.The possible misregister of user's refund information, exists
The case where information is put in storage not in time.Therefore, can automate inspection user refund information whether vacancy, completion refund information
Deng.
Data correctness verification for example, refund in the data that verifying refund date amendment is completed date and refund state
Whether rationally, it verifies date amendment of refunding and completes the accuracy that the loan of data counts by stages, it is ensured that each only has one by stages
Data.It such as finds there are problems that loan repeats to record by stages in data, duplicate information will be removed.
History refund Information Statistics for example, can be to each in the data correction verification for completing front
Data are grouped according to the loan odd numbers belonging to it, i.e., form one group for each data by stages of same loan, dividing
The statistics that history refund information is carried out in group, generates history refund statistical information.
Data normalization for example, Missing Data Filling, normalization onehot coding are done to the data that aforementioned processing is completed
Deng.
Predictive information can be presented using forms such as tables, for example, overdue of that month prediction statistical form is obtained, in loan
Overdue data time month predict statistical form, are predicted after overdue data 2 months in statistical form, or the loan of the longer time following phase in loan
Overdue prediction statistical form.
Fig. 3 is the structural schematic diagram one of the refund information prediction device of loan provided by the embodiments of the present application, such as Fig. 3 institute
Show, the device of the present embodiment may include: to obtain module 31 and prediction module 32；
Module is obtained, for by stages corresponding described according to the first of loan the first refund information by stages and described first
First history refund information of loan, the second history for obtaining the second of the loan the corresponding loan by stages are refunded letter
Breath；Described first by stages for second by stages upper one by stages；The first refund information is the picture according to the user of the loan
As information, the first history refund information and machine learning model trained in advance, the user predicted are directed to
Described first refund information by stages；
Prediction module is used for according to the portrait information, the second history refund information and the machine learning model,
Predict the user for described second the second refund information by stages.
The device of the present embodiment can be used for executing the technical solution of above method embodiment, realization principle and technology
Effect is similar, and details are not described herein again.
In a kind of possible design, the prediction module 32 is specifically used for:
By machine learning model described in the portrait information, the second history refund information input, a pre- direction finding is exported
Amount, the predicted vector are used to indicate the second refund information.
In a kind of possible design, the component of the predicted vector include the user for described second by stages also
The overdue number of days of money is located at each probability preset in overdue section；
If the maximum probability in each probability is greater than preset threshold, the second refund information includes the user for institute
It states the overdue number of days of the second refund by stages and is located at target and preset in overdue section；Wherein, the target is preset described in overdue section
Maximum probability is corresponding to preset overdue section.
In a kind of possible design, the first history refund information includes the described first corresponding loan by stages
The first history refund statistical information, the second history refund statistical information includes the described second corresponding loan by stages
The second history refund statistical information；The acquisition module 31, is specifically used for:
Overdue section and the first history refund statistical information are preset according to the corresponding target of the first refund information,
Obtain the second history refund statistical information；
According to the second history refund statistical information, the second history refund information is obtained.
In a kind of possible design, the first history refund information includes: the described first corresponding loan by stages
The refund information of the history of money each phase；The second history refund information includes: the described second corresponding loan by stages
The refund information of history each phase；The acquisition module 31, is specifically used for:
Overdue section and the described first corresponding loan by stages are preset according to the corresponding target of the first refund information
The refund information of the history of money each phase obtains the refund information of described second history each phase of the corresponding loan by stages；
According to the refund information of described second history each phase of the corresponding loan by stages, second history is obtained also
Money information.
In a kind of possible design, the first history refund information includes: the described first corresponding loan by stages
The refund information of first history refund statistical information of money and described first history each phase of the corresponding loan by stages, it is described
Second history refund information includes the described second second history refund statistical information of the corresponding loan and described by stages
The refund information of two history each phases of the corresponding loan by stages；The acquisition module 31, is specifically used for:
Overdue section and the first history refund statistical information are preset according to the corresponding target of the first refund information, is obtained
Second history refund statistical information；
Overdue section and the described first corresponding loan by stages are preset according to the corresponding target of the first refund information
The refund information of the history of money each phase obtains the refund information of described second history each phase of the corresponding loan by stages；
According to the second history refund statistical information and described second history each phase of the corresponding loan by stages
Refund information obtains the second history refund information.
The device of the present embodiment can be used for executing the technical solution of above method embodiment, realization principle and technology
Effect is similar, and details are not described herein again.
Fig. 4 is the structural schematic diagram two of the refund information prediction device of loan provided by the embodiments of the present application, such as Fig. 4 institute
Show, can also include: training module 33 further on the basis of the device of the present embodiment apparatus structure shown in Fig. 3；Institute
Training module 33 is stated to be used for:
Multiple training samples are obtained, the training sample includes that once providing a loan for training user is each by stages corresponding
The information of training user, the information of the training user include that the portrait information of the training user is gone through with accordingly by stages corresponding
History refund information；
Obtain the label of each training sample, the label be used to indicate corresponding training sample it is corresponding once provide a loan it is each
Corresponding refund information by stages；
The training of the multiple training sample is obtained according to the label of the multiple training sample and each training sample
The machine learning model.
The device of the present embodiment can be used for executing the technical solution of above method embodiment, realization principle and technology
Effect is similar, and details are not described herein again.
Fig. 5 is the structural schematic diagram of electronic equipment provided by the embodiments of the present application；Referring to Fig. 5, the electronics of the present embodiment is set
Standby 500 include: processor 51 and memory 52, and the processor 51 is coupled with memory 52；
The memory 52 is used for, and stores computer program；
The processor 51 is used for, and executes the computer program stored in the memory 52, so that the electronics is set
It is standby to execute method described in above-mentioned any means embodiment.
In a kind of mode, memory 52 can be the memory outside electronic equipment 500.
The embodiment of the present application provides a kind of readable storage medium storing program for executing, including program or instruction, when described program or instruction are being counted
When running on calculation machine, the method as described in above-mentioned any means embodiment is performed.
Those of ordinary skill in the art will appreciate that: realize that all or part of the steps of above-mentioned each method embodiment can lead to
The relevant hardware of program instruction is crossed to complete.Program above-mentioned can be stored in a computer readable storage medium.The journey
When being executed, execution includes the steps that above-mentioned each method embodiment to sequence；And storage medium above-mentioned include: ROM, RAM, magnetic disk or
The various media that can store program code such as person's CD.
Finally, it should be noted that the above various embodiments is only to illustrate the technical solution of the embodiment of the present application, rather than to it
Limitation；Although the embodiment of the present application is described in detail referring to foregoing embodiments, those skilled in the art
It is understood that it is still possible to modify the technical solutions described in the foregoing embodiments, either to part of or
All technical features are equivalently replaced；And these are modified or replaceed, it does not separate the essence of the corresponding technical solution this Shen
Please example scheme range.
Claims (12)
1. a kind of refund information forecasting method of loan characterized by comprising
According to the first of loan the first refund information by stages and described first by stages the corresponding loan the first history also
Money information obtains the second history refund information of the second of the loan the corresponding loan by stages；Described first is by stages
Described second by stages upper one by stages；The first refund information is according to the portrait information of the user of the loan, described
One history refund information and machine learning model trained in advance, the user predicted for described first by stages
Refund information；
According to the portrait information, the second history refund information and the machine learning model, predict that the user is directed to
Described second the second refund information by stages.
2. the method according to claim 1, wherein it is described according to the portrait information, second history also
Money information and the machine learning model predict the user for described second the second refund information by stages, comprising:
By machine learning model described in the portrait information, the second history refund information input, predicted vector is exported, it is described
Predicted vector is used to indicate the second refund information.
3. according to the method described in claim 2, it is characterized in that, the component of the predicted vector includes the user for institute
It states the second overdue number of days of refund by stages and is located at each probability preset in overdue section；
If maximum probability in each probability is greater than preset threshold, the second refund information includes the user for described the
The overdue number of days of two refund by stages, which is located at target, to be preset in overdue section；Wherein, the target preset overdue section be it is described most
Maximum probability is corresponding to preset overdue section.
4. according to the method described in claim 3, it is characterized in that, the first history refund information includes: described first point
First history refund statistical information of the phase corresponding loan, the second history refund statistical information include described second point
Second history refund statistical information of the phase corresponding loan；The first first refund information by stages according to loan and
First history refund information of the described first corresponding loan by stages, obtain the loan second are by stages corresponding described
Second history refund information of loan, comprising:
Overdue section and the first history refund statistical information are preset according to the corresponding target of the first refund information, is obtained
Second history refund statistical information；
According to the second history refund statistical information, the second history refund information is obtained.
5. according to the method described in claim 3, it is characterized in that, the first history refund information includes: described first point
The refund information of history each phase of the phase corresponding loan；The second history refund information includes: described second by stages right
The refund information of the history for the loan answered each phase；The first first refund information and described by stages according to loan
First history refund information of the one corresponding loan by stages obtains the second of the loan the corresponding loan by stages
Second history refund information, comprising:
Overdue section and the described first corresponding loan by stages are preset according to the first refund information corresponding target
The refund information of history each phase obtains the refund information of described second history each phase of the corresponding loan by stages；
According to the refund information of described second history each phase of the corresponding loan by stages, obtains second history and refund letter
Breath.
6. according to the method described in claim 3, it is characterized in that, the first history refund information includes: described first point
History each phase of first history refund statistical information of the phase corresponding loan and the described first corresponding loan by stages
Refund information, the second history refund information include described second by stages the corresponding loan the second history refund system
Count the refund information of information and described second history each phase of the corresponding loan by stages；It is described according to the first of loan by stages
The first refund information and the described first corresponding loan by stages the first history refund information, obtain the of the loan
Second history refund information of two corresponding loans by stages, comprising:
Overdue section and the first history refund statistical information are preset according to the corresponding target of the first refund information, obtains second
History refund statistical information；
Overdue section and the described first corresponding loan by stages are preset according to the first refund information corresponding target
The refund information of history each phase obtains the refund information of described second history each phase of the corresponding loan by stages；
According to the refund of the second history refund statistical information and described second history each phase of the corresponding loan by stages
Information obtains the second history refund information.
7. the method according to claim 1, which is characterized in that the method also includes:
Multiple training samples are obtained, the training sample includes each by stages corresponding training of training user once provided a loan
The information of user, the portrait information and corresponding by stages corresponding history that the information of the training user includes the training user are also
Money information；
Obtain the label of each training sample, the label be used to indicate corresponding training sample it is corresponding once provide a loan it is each by stages
Corresponding refund information；
The training of the multiple training sample is obtained described according to the label of the multiple training sample and each training sample
Machine learning model.
8. the method according to the description of claim 7 is characterized in that the training user is to match with the user, and borrow
The user that money type and the type of the loan match.
9. a kind of refund information prediction device of loan characterized by comprising
Module is obtained, for according to the first of loan the first refund information by stages and the described first corresponding loan by stages
The first history refund information, obtain the second history refund information of the second of the loan the corresponding loan by stages；Institute
State first by stages for second by stages upper one by stages；The first refund information is believed according to the portrait of the user of the loan
Breath, the first history refund information and machine learning model trained in advance, the user predicted is for described
First refund information by stages；
Prediction module, for according to the portrait information, the second history refund information and the machine learning model, prediction
The user is directed to described second the second refund information by stages.
10. device according to claim 9, which is characterized in that further include training module；The training module is used for:
Multiple training samples are obtained, the training sample includes each by stages corresponding training of training user once provided a loan
The information of user, the portrait information and corresponding by stages corresponding history that the information of the training user includes the training user are also
Money information；
Obtain the label of each training sample, the label be used to indicate corresponding training sample it is corresponding once provide a loan it is each by stages
Corresponding refund information；
The training of the multiple training sample is obtained described according to the label of the multiple training sample and each training sample
Machine learning model.
11. a kind of readable storage medium storing program for executing, which is characterized in that including program or instruction, when described program or instruct on computers
When operation, any method of claim 1~8 is performed.
12. a kind of electronic equipment characterized by comprising processor, the processor are coupled with memory；
The memory is used for, and stores computer program；
The processor is used for, and calls the computer program stored in the memory, to realize any institute of claim 1~8
The method stated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811151423.8A CN109242672A (en) | 2018-09-29 | 2018-09-29 | Refund information forecasting method, device and the computer readable storage medium of loan |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811151423.8A CN109242672A (en) | 2018-09-29 | 2018-09-29 | Refund information forecasting method, device and the computer readable storage medium of loan |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109242672A true CN109242672A (en) | 2019-01-18 |
Family
ID=65054157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811151423.8A Pending CN109242672A (en) | 2018-09-29 | 2018-09-29 | Refund information forecasting method, device and the computer readable storage medium of loan |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109242672A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109859032A (en) * | 2019-01-22 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Funds on account collection method, apparatus, equipment and storage medium based on intelligent sound |
| CN110322341A (en) * | 2019-06-24 | 2019-10-11 | 阿里巴巴集团控股有限公司 | The determination method and apparatus of credit label |
| CN110363658A (en) * | 2019-07-16 | 2019-10-22 | 北京明略软件系统有限公司 | Processing method and processing device, storage medium and the electronic device of credit data |
| CN110782338A (en) * | 2019-10-15 | 2020-02-11 | 卓尔智联(武汉)研究院有限公司 | Loan transaction risk prediction method and device, computer equipment and storage medium |
| CN110826921A (en) * | 2019-11-08 | 2020-02-21 | 腾讯科技（深圳）有限公司 | Data processing method, data processing device, computer readable storage medium and computer equipment |
| CN111192131A (en) * | 2019-12-12 | 2020-05-22 | 上海淇玥信息技术有限公司 | Financial risk prediction method and device and electronic equipment |
| CN112634018A (en) * | 2020-12-23 | 2021-04-09 | 百维金科(上海)信息科技有限公司 | Overdue monitoring method for optimizing recurrent neural network based on ant colony algorithm |
| WO2021086814A1 (en) * | 2019-11-01 | 2021-05-06 | Square, Inc. | System and method for generating dynamic repayment terms |
| CN112991052A (en) * | 2021-04-25 | 2021-06-18 | 大箴(杭州)科技有限公司 | Repayment capability evaluation method and device |
| CN113177837A (en) * | 2021-05-12 | 2021-07-27 | 广州市全民钱包科技有限公司 | Loan amount evaluation method, device, equipment and storage medium for loan applicant |
| CN113313569A (en) * | 2021-05-18 | 2021-08-27 | 建信金融科技有限责任公司 | Method, device, equipment and storage medium for predicting repayment amount |
| CN114971879A (en) * | 2022-06-17 | 2022-08-30 | 北京极致车网科技有限公司 | Information processing system and information processing method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170061534A1 (en) * | 2015-08-28 | 2017-03-02 | Mastercard International Incorporated | Assessing credit risk |
| CN106909981A (en) * | 2015-12-23 | 2017-06-30 | 阿里巴巴集团控股有限公司 | Model training, sample balance method and device and personal credit points-scoring system |
| CN107993142A (en) * | 2017-11-23 | 2018-05-04 | 搜易贷（北京）金融信息服务有限公司 | A kind of anti-risk of fraud control system of finance |
| CN108256667A (en) * | 2017-12-04 | 2018-07-06 | 重庆金融资产交易所有限责任公司 | Asset data processing method, device, storage medium and computer equipment |
-
2018
- 2018-09-29 CN CN201811151423.8A patent/CN109242672A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170061534A1 (en) * | 2015-08-28 | 2017-03-02 | Mastercard International Incorporated | Assessing credit risk |
| CN106909981A (en) * | 2015-12-23 | 2017-06-30 | 阿里巴巴集团控股有限公司 | Model training, sample balance method and device and personal credit points-scoring system |
| CN107993142A (en) * | 2017-11-23 | 2018-05-04 | 搜易贷（北京）金融信息服务有限公司 | A kind of anti-risk of fraud control system of finance |
| CN108256667A (en) * | 2017-12-04 | 2018-07-06 | 重庆金融资产交易所有限责任公司 | Asset data processing method, device, storage medium and computer equipment |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109859032A (en) * | 2019-01-22 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Funds on account collection method, apparatus, equipment and storage medium based on intelligent sound |
| CN110322341A (en) * | 2019-06-24 | 2019-10-11 | 阿里巴巴集团控股有限公司 | The determination method and apparatus of credit label |
| CN110322341B (en) * | 2019-06-24 | 2024-01-05 | 创新先进技术有限公司 | Method and device for determining credit label |
| CN110363658A (en) * | 2019-07-16 | 2019-10-22 | 北京明略软件系统有限公司 | Processing method and processing device, storage medium and the electronic device of credit data |
| CN110782338A (en) * | 2019-10-15 | 2020-02-11 | 卓尔智联(武汉)研究院有限公司 | Loan transaction risk prediction method and device, computer equipment and storage medium |
| US11145005B2 (en) | 2019-11-01 | 2021-10-12 | Square, Inc. | System and method for generating dynamic repayment terms |
| US11348190B2 (en) | 2019-11-01 | 2022-05-31 | Block, Inc. | System and method for generating dynamic repayment terms |
| WO2021086814A1 (en) * | 2019-11-01 | 2021-05-06 | Square, Inc. | System and method for generating dynamic repayment terms |
| CN110826921A (en) * | 2019-11-08 | 2020-02-21 | 腾讯科技（深圳）有限公司 | Data processing method, data processing device, computer readable storage medium and computer equipment |
| CN110826921B (en) * | 2019-11-08 | 2023-04-18 | 腾讯科技（深圳）有限公司 | Data processing method, data processing device, computer readable storage medium and computer equipment |
| CN111192131A (en) * | 2019-12-12 | 2020-05-22 | 上海淇玥信息技术有限公司 | Financial risk prediction method and device and electronic equipment |
| CN112634018A (en) * | 2020-12-23 | 2021-04-09 | 百维金科(上海)信息科技有限公司 | Overdue monitoring method for optimizing recurrent neural network based on ant colony algorithm |
| CN112991052A (en) * | 2021-04-25 | 2021-06-18 | 大箴(杭州)科技有限公司 | Repayment capability evaluation method and device |
| CN113177837A (en) * | 2021-05-12 | 2021-07-27 | 广州市全民钱包科技有限公司 | Loan amount evaluation method, device, equipment and storage medium for loan applicant |
| CN113313569A (en) * | 2021-05-18 | 2021-08-27 | 建信金融科技有限责任公司 | Method, device, equipment and storage medium for predicting repayment amount |
| CN114971879A (en) * | 2022-06-17 | 2022-08-30 | 北京极致车网科技有限公司 | Information processing system and information processing method |
| CN114971879B (en) * | 2022-06-17 | 2023-04-07 | 北京极致车网科技有限公司 | Information processing system and information processing method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109242672A (en) | Refund information forecasting method, device and the computer readable storage medium of loan | |
| CN108846520A (en) | Overdue loan prediction technique, device and computer readable storage medium | |
| CN111192131A (en) | Financial risk prediction method and device and electronic equipment | |
| US11645656B2 (en) | Machine learning-based determination of limits on merchant use of a third party payments system | |
| CN110827138A (en) | Push information determining method and device | |
| US11687936B2 (en) | System and method for managing chargeback risk | |
| CN108492170A (en) | A kind of repayment method, computer readable storage medium and terminal device | |
| US20160042354A1 (en) | Predictive fraud screening | |
| AU2021225802A1 (en) | Machine-learning techniques to generate recommendations for risk mitigation | |
| US10282781B2 (en) | Data analytics database and platform system and method | |
| CN111144899B (en) | Method and device for identifying false transaction and electronic equipment | |
| US20140025454A1 (en) | Early Payment Rewards System and Method | |
| CN103366270A (en) | Multi-platform data interaction method and system | |
| KR102174608B1 (en) | Apparatus for predicting loan defaults based on machine learning and method thereof | |
| CN108229804A (en) | Brand promotion project management method, device, terminal device and storage medium | |
| CN114971893B (en) | Loan deduction mode flexible configuration method and system | |
| CN108256667A (en) | Asset data processing method, device, storage medium and computer equipment | |
| US9633357B2 (en) | Net utility determination based on product replacement and service plan coverage decisions | |
| CN111899093B (en) | Method and device for predicting default loss rate | |
| CN112465611A (en) | Method and device for pushing article information to user and electronic equipment | |
| CN110705975B (en) | Deduction instruction initiating method and device | |
| CN108416556B (en) | Cargo attention balancing method for vehicle-cargo matching platform | |
| CN109472704A (en) | Screening technique, terminal device and the medium of fund product neural network based | |
| CN114240610B (en) | Automatic fund collection method, device, computer equipment and storage medium | |
| CN112116169B (en) | User behavior determining method and device and electronic equipment |
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 |