CN110147938A - A kind of training sample generation method, device, system and recording medium - Google Patents
A kind of training sample generation method, device, system and recording medium Download PDFInfo
- Publication number
- CN110147938A CN110147938A CN201910330919.XA CN201910330919A CN110147938A CN 110147938 A CN110147938 A CN 110147938A CN 201910330919 A CN201910330919 A CN 201910330919A CN 110147938 A CN110147938 A CN 110147938A
- Authority
- CN
- China
- Prior art keywords
- value
- user
- user characteristics
- sample
- knee
- 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
- 238000012549 training Methods 0.000 title claims abstract description 44
- 238000000034 method Methods 0.000 title claims abstract description 23
- 210000003127 knee Anatomy 0.000 claims abstract description 48
- 239000000284 extract Substances 0.000 claims abstract description 7
- 238000007405 data analysis Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000032683 aging Effects 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 230000003542 behavioural effect Effects 0.000 description 3
- 230000007812 deficiency Effects 0.000 description 3
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000009432 framing Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Marketing (AREA)
- Data Mining & Analysis (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Business, Economics & Management (AREA)
- Game Theory and Decision Science (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Technology Law (AREA)
- Operations Research (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention discloses a kind of training sample generation method, device, equipment and computer-readable mediums based on user characteristics knee value.The method includes obtaining user's sample and extract at least one user characteristics；Judged according to the knee value of at least one user characteristics, to determine that user's sample attribute, the attribute include first property value and the second attribute value；And model training is carried out according to user's sample of determining user's sample attribute.The present invention can shorten the samples selection time, improve the experience of Debit User on the basis of guaranteeing sample accuracy.
Description
Technical field
The present invention relates to technical field of internet application, and in particular to a kind of training sample for credit scoring model is raw
At method, apparatus, system and recording medium.
Background technique
In recent years, the further application with machine learning techniques in reference field, is applied in credit investigation system at model
It is high to the degree of dependence of training data in diversification, the stage of syncretization, it can preferably meet mechanism commenting for user credit
Valence demand.It is now to serve all kinds of internet credit information services, lending institution more, using it is more be credit scoring system.
Credit scoring system chooses user's sample usually from known users, according to the behavioral data of user's sample and user
Attribute data can extract the user characteristics of user's sample, be instructed by the user characteristics of user's sample to Rating Model
Practice, credit scoring is carried out to user using trained Rating Model.
In the implementation of the present invention, the inventor finds that the existing technology has at least the following problems:
When establishing personal credit Rating Model, the spy of enough characterization credit applications people behavior of credit is not only needed
Variable is levied, and the capacity of modeling sample must also reach certain quantity.In general, sample size is bigger, is established
The precision or predictive ability of model are higher, and model is also more steady, therefore often face the problem of client's sample deficiency.
In addition, in certain scenes, accumulation sample data and training pattern, thus realize the deployment of machine learning model,
Longer time is generally required, is set to as the observation period of air control model and performance phase 1 year or more, application is caused to score
Card mold type has natural hysteresis quality, therefore brings very big influence to the assessment of user credit.
Based on the prior art, more efficient, more accurate model training sample is needed.
Summary of the invention
In view of the foregoing deficiencies of prior art, the purpose of the present invention is to provide a kind of efficient, accurate user's instructions
Practice sample generating method.
In order to solve the above technical problems, the first aspect of the present invention proposes a kind of training based on user characteristics knee value
Sample generating method includes the following steps: to obtain user's sample and extracts at least one user characteristics；According at least one user
The knee value of feature is judged, to determine that user's sample attribute, the attribute include first property value and the second attribute value；Root
Model training is carried out according to user's sample of determining user's sample attribute.
A preferred embodiment of the invention, the method also includes inflection point occur at least one user characteristics
The time point of value carries out sample attribute judgement.
A preferred embodiment of the invention, the knee value are based at least one user characteristics and carry out big data
Analysis obtains.Equidistant observation point is specially set within the performance phase；At least one user characteristics is obtained in Current observation point
Value；Judge whether Current observation value and the difference of previous observation meet preset condition；When meeting preset condition, Current observation point
Value be knee value.
A preferred embodiment of the invention, the knee value are based at least one user characteristics and carry out big data
Analysis obtains.The inflection point for specially storing at least one user characteristics goes out the empirical value that the characteristic value of current moment changes；According to pre-
Fixed cycle obtains the value of at least one user characteristics；And the empirical value for inflection point going out the characteristic value of current moment is compared, and is judged
Whether the current time reaches inflection point.
A preferred embodiment of the invention, the user characteristics include at least one of following data: overdue
Rate, amount owed, debt duration, overdue repayment amount, overdue refund duration.
A preferred embodiment of the invention, first property value 1 represent hospitable family；Second attribute value is 0,
Represent bad client.
The second aspect of the present invention proposes a kind of training sample generating means based on user characteristics knee value, comprising: obtains
Modulus block obtains user's sample and extracts at least one user characteristics；Determining module, according to the inflection point of at least one user characteristics
Value is judged, to determine that user's sample attribute, the attribute include the first value and second value；Training module, according to described true
The user's sample for determining attribute carries out model training.
A preferred embodiment of the invention, described device further include judgment module, at least one use
The time point that knee value occurs in family feature carries out sample attribute judgement.
A preferred embodiment of the invention, described device further includes that inflection point obtains module, for based at least
One user characteristics carries out big data analysis and obtains the knee value, comprising:
Equidistant observation point is arranged in setup module within the performance phase；
Acquisition module obtains at least one user characteristics in the value of Current observation point；
Generation module, judges whether Current observation value and the difference of previous observation meet preset condition；Meet default item
When part, the value of Current observation point is knee value.
A preferred embodiment of the invention, described device further includes that inflection point obtains module, for based at least
One user characteristics carries out big data analysis and obtains the knee value, comprising:
Empirical data memory module, the inflection point for storing at least one user characteristics go out the experience that the characteristic value of current moment changes
Value；
Characteristic information obtains module, and the value of at least one user characteristics is obtained according to predetermined period；
Inflection point judgment module, and the empirical value for going out the characteristic value of current moment with inflection point is compared, when judging described current
It carves and whether reaches inflection point.
A preferred embodiment of the invention, the user characteristics include at least one of following data: overdue
Rate, amount owed, debt duration, overdue repayment amount, overdue refund duration.
A preferred embodiment of the invention, first property value 1 represent hospitable family；Second attribute value is 0,
Represent bad client.
The third aspect of the present invention proposes a kind of training sample generating device based on user characteristics knee value, comprising: deposits
Reservoir, for storing computer executable program；Data processing equipment, it is executable for reading the computer in the memory
Program, to execute the training sample generation method based on user characteristics knee value.
The fourth aspect of the present invention proposes a kind of computer-readable medium, for storing computer-readable program, the meter
Calculation machine readable program is for executing the training sample generation method based on user characteristics knee value.
The present invention used within the performance phase, the time point that at least one user characteristics knee value of real-time tracking occurs, at this
Time point carries out sample attribute and then using the user's sample progress model training for determining user's sample attribute, solves current use
The problem of family sample data lacks, training pattern lags.Training sample generation side based on user characteristics knee value of the invention
Method can shorten the samples selection time, improve the experience of Debit User on the basis of guaranteeing sample accuracy.
Detailed description of the invention
Fig. 1 is the flow diagram of the training sample of the invention based on user characteristics knee value；
Fig. 2 is overdue rate distribution map in the performance phase provided in an embodiment of the present invention；
Fig. 3 (a), Fig. 3 (b) are the module architectures of the training sample generating means of the invention based on user characteristics knee value
Schematic diagram；
Fig. 4 is the structural framing schematic diagram of the training sample generating device of the invention based on user characteristics knee value.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with specific embodiment, and reference
Attached drawing, the present invention is described in further detail.
Fig. 1 is the flow diagram of the training sample generation method the present invention is based on user characteristics knee value.Such as Fig. 1 institute
Show, method of the invention has following steps:
S1, user's sample and at least one user characteristics is extracted firstly, obtaining.
Original user data is obtained from user data source first, original user data can be obtained from external data source
It takes, can also be obtained from internal data source, or obtained jointly in conjunction with external data source with internal data source.Wherein, the use
User data may include the data such as the attribute data, behavioral data and credit record of user.Original user data is converted into uniting
One format deletes abnormal data and missing data therein, generates user's sample data.
Each user characteristics are for describing the information of user's sample data in one aspect.For example, in fiduciary loan business
In, when the user characteristics of user's sample data may include: amount owed, debt duration, overdue repayment amount, overdue refund
Long, overdue rate etc..
Identified user characteristics can be specified directly, alternatively, can also be by preset rules from original training sample packet
It is selected in the multiple characteristic variables included.Wherein, preset rules can be determined according to the operation target of practical business scene, or
Person, model training direction determine.By taking Credit Risk Model as an example, the target of Credit Risk Model is to identify to be kept outside of the door
Potential bad client, therefore provide the basis that the classification of reasonable client's quality is modeling.For example, in credit scoring field, generally
Portray the behavior of client with the overdue parameter of client, the more bad client of overdue issue, the overdue amount of money, undesirable level is higher.
In this way, at least one user characteristics determined by step S1 may include at least one of following data: overdue rate, debt
The amount of money, debt duration, overdue repayment amount, overdue refund duration etc..
Wherein, this can not be limited for one or more, the application in the extracted user characteristics of step S1.It needs
Illustrate, when identified user characteristics are multiple, is individually handled for the user characteristics of each determination.
S2, judged according to the knee value of at least one user characteristics, to determine user's sample attribute, the attribute packet
Include first property value and the second attribute value.
In the present invention, user's sample attribute is defined as comprising two class of first property value and the second attribute value, wherein first
Attribute value is 1, represents user's sample preferably client, and the second attribute value is 0, represents user's sample as bad client.Credit risk mould
Type is briefly exactly to grab the feature that bad client is substantially distinguished from hospitable family by historical data, and go as standard pre-
Survey future has the people that this behavior occurs in very maximum probability.So the definition of " good ", " bad " is just particularly important in a model.
Common practice be setting the performance phase, within the performance phase collect customers' credit performance user characteristics, such as whether it is overdue, whether drag
Owe, income height it is low, to ensure that the performance phase setting of client meets present business actual conditions, and client can be made as far as possible
Risk sufficiently shows, and shows the phase by the way of rolling time length, so that the performance phase of user's sample is not less than 12 months.
Within the performance phase, user characteristics value (Y-axis) changes with the variation of time (X-axis), and shows certain rule
Rule property, as shown in Fig. 2, user characteristics value is suddenly substantially increased with time series, tends towards stability after undergoing a period of time.It turns
Point then refers to that time series becomes the point to tend towards stability from ascendant trend, and knee value is to go out the user characteristics of current moment in inflection point
Value.
User characteristics are by taking overdue rate as an example in the present embodiment, in credit operation, in overdue number of days greater than 60 days
Client is difficult collection substantially, i.e., client of the overdue number of days greater than 60 days is essentially bad client.By aging (Vintage) point
Analysis, earlier month of the overdue rate of the odd corpus monthly made loans overdue 60 days or more after loan examination & approval pass through suddenly rise, base
This tends towards stability when aging is 8, i.e., the probability that hospitable family becomes bad client when aging is greater than 8 substantially reduces, i.e. the credit
The inflection point (points of circle markings in Fig. 2) of business appears in the time point that aging is 8, therefore the sample performance phase is only covered the
8 phases, then can define sample characteristics variable is that client of the overdue number of days greater than 60 days is bad client within 8 phases, remaining visitor
Family preferably client.This has the difference of very little with 1 year or more user's sample will be arranged in the phase of performance.
Therefore, in order to solve the technical problem due to performance phase too long brought information delay and sample deficiency, by table
Current length is set as the time of knee value appearance, carries out the judgement of sample attribute at the time point, according to established rule, really
Random sample originally belongs to hospitable family or bad client, promotes assessment efficiency.
However according to business actual conditions, the time point that inflection point occurs also is not quite similar, therefore using based on user characteristics
The mode for carrying out big data analysis obtains accurate inflection point information.
Mode in the present embodiment is that equidistant observation point is arranged within the performance phase；At least one user characteristics is obtained to exist
The value of Current observation point；Judge whether Current observation value and the difference of previous observation meet preset condition；Meet preset condition
When, the value of Current observation point is knee value.Wherein, equidistant observation point can be using the moon as the period, when inflection point occurs, front and back
The difference of observation point is substantially reduced, based on experience value given threshold, is met threshold condition and is met preset condition.
Alternatively, the inflection point for storing at least one user characteristics goes out the empirical value of the characteristic value variation of current moment；According to predetermined
Period obtains the value of at least one user characteristics；And the empirical value for inflection point going out the characteristic value of current moment is compared, and judges institute
State whether current time reaches inflection point.Predetermined period can be the period moon.
S3, finally, according to user's sample of determining user's sample attribute carry out model training.
The present embodiment can be user's sample according to preset strategy after the user's sample for obtaining determining user's sample attribute
Then this setting weight is trained default assessment models using user's sample, assessment models after being trained, and be based on
Assessment models assess the internet reference of user after the training.
Fig. 3 (a), 3 (b) be training sample generating means the present invention is based on user characteristics knee value configuration diagram.
As shown in figure 4, the inventive system comprises obtain module, determining module and training module.Generally speaking, module is obtained for obtaining
It takes family sample and extracts at least one user characteristics.
Original user data is obtained from user data source first, original user data can be obtained from external data source
It takes, can also be obtained from internal data source, or obtained jointly in conjunction with external data source with internal data source.Wherein, the use
User data may include the data such as the attribute data, behavioral data and credit record of user.Original user data is converted into uniting
One format deletes abnormal data and missing data therein, generates user's sample data.Identified user characteristics can be straight
Connect it is specified, alternatively, can also be selected from multiple characteristic variables that original training sample includes by preset rules.Institute is really
When at least one fixed user characteristics may include overdue rate, amount owed, debt duration, overdue repayment amount, overdue refund
It is long etc..
Secondly, determining module is then used to be judged according to the knee value of at least one user characteristics, to determine user's sample
This attribute, the attribute include first property value and the second attribute value.
Determining module also includes judgment module, carries out sample for occurring the time point of knee value at least one user characteristics
This determined property；And inflection point obtains module and is used to be based on user characteristics progress big data analysis at least one to obtain described turn
Point value.Mode in the present embodiment is that equidistant observation point is arranged within the performance phase；At least one user characteristics is obtained current
The value of observation point；Judge whether Current observation value and the difference of previous observation meet preset condition；When meeting preset condition, when
The value of preceding observation point is knee value.
Alternatively, the inflection point for storing at least one user characteristics goes out the empirical value of the characteristic value variation of current moment；According to predetermined
Period obtains the value of at least one user characteristics；And the empirical value for inflection point going out the characteristic value of current moment is compared, and judges institute
State whether current time reaches inflection point.
Finally, training module is used to carry out model training according to user's sample of determining user's sample attribute.
In addition, the present invention also proposes a kind of training sample generating device based on user characteristics knee value.Fig. 4 is the present invention
The training sample generating device based on user characteristics knee value structural framing schematic diagram, as shown in figure 4, the equipment include deposit
Reservoir and data processing equipment, memory is for storing computer executable program, data processing equipment, for reading described deposit
Computer executable program in reservoir, to execute the training sample generation method based on user characteristics knee value.This
The memory of invention can be local storage, be also possible to distributed memory system, such as cloud storage system.And data processing
Device then includes the device of at least one tool number word information processing capability, such as CPU, GPU, multicomputer system or cloud processing
Device.
Furthermore the present invention also proposes computer-readable medium, described computer-readable for storing computer-readable program
Program is used to execute the training sample generation method based on user characteristics knee value.
It should be appreciated that in order to simplify the present invention and help it will be understood by those skilled in the art that various aspects of the invention,
Above in the description of exemplary embodiment of the present invention, each feature of the invention is retouched in a single embodiment sometimes
It states, or is described referring to single figure.But should not be by the feature that the present invention is construed to include in exemplary embodiment
The essential features of patent claims.
It should be appreciated that can be to progress such as module, unit, the components for including in the equipment of one embodiment of the present of invention certainly
It adaptively changes so that they are arranged in equipment unlike this embodiment.The difference that can include the equipment of embodiment
Module, unit or assembly are combined into module, a unit or assembly, also they can be divided into multiple submodule, subelement or
Sub-component.Module, unit or assembly in the embodiment of the present invention can realize in hardware, can also be with one or more
The software mode run on a processor is realized, or is implemented in a combination thereof.
Particular embodiments described above has carried out further in detail the purpose of the present invention, technical scheme and beneficial effects
Describe in detail bright, it should be understood that the above is only a specific embodiment of the present invention, is not intended to restrict the invention, it is all
Within the spirit and principles in the present invention, any modification, equivalent substitution, improvement and etc. done should be included in protection of the invention
Within the scope of.
Claims (10)
1. a kind of training sample generation method based on user characteristics knee value, includes the following steps:
It obtains user's sample and extracts at least one user characteristics；
Judged according to the knee value of at least one user characteristics, to determine that user's sample attribute, the attribute include first
Attribute value and the second attribute value；
Model training is carried out according to user's sample of determining user's sample attribute.
2. the method as described in claim 1, it is characterised in that:
Sample attribute judgement is carried out at the time point that knee value occurs at least one user characteristics.
3. method according to claim 1 or 2, which is characterized in that the knee value be based at least one user characteristics into
Row big data analysis obtains, comprising:
Equidistant observation point is set within the performance phase；
At least one user characteristics is obtained in the value of Current observation point；
Judge whether Current observation value and the difference of previous observation meet preset condition；
When meeting preset condition, the value of Current observation point is knee value.
4. method according to claim 1 or 2, which is characterized in that the knee value be based at least one user characteristics into
Row big data analysis obtains, comprising:
The inflection point for storing at least one user characteristics goes out the empirical value that the characteristic value of current moment changes；
The value of at least one user characteristics is obtained according to predetermined period；
And the empirical value for inflection point going out the characteristic value of current moment is compared, and judges whether the current time reaches inflection point.
5. a kind of training sample generating means based on user characteristics knee value characterized by comprising
Module is obtained, user's sample is obtained and extracts at least one user characteristics；
Determining module is judged according to the knee value of at least one user characteristics, to determine user's sample attribute, the attribute
Including the first value and second value；
Training module carries out model training according to user's sample of the determining attribute.
6. device according to claim 5, which is characterized in that determining module also includes judgment module, at least one use
The time point that knee value occurs in family feature carries out sample attribute judgement.
7. device according to claim 5 or 6, which is characterized in that determining module also includes that inflection point obtains module, is used for base
Big data analysis, which is carried out, at least one user characteristics obtains the knee value, comprising:
Equidistant observation point is arranged in setup module within the performance phase；
Acquisition module obtains at least one user characteristics in the value of Current observation point；
Generation module, judges whether Current observation value and the difference of previous observation meet preset condition；When meeting preset condition,
The value of Current observation point is knee value.
8. device according to claim 5 or 6, which is characterized in that determining module also includes that inflection point obtains module, is used for base
Big data analysis, which is carried out, at least one user characteristics obtains the knee value, comprising:
Empirical data memory module, the inflection point for storing at least one user characteristics go out the empirical value that the characteristic value of current moment changes；
Characteristic information obtains module, and the value of at least one user characteristics is obtained according to predetermined period；
Inflection point judgment module, and the empirical value for going out the characteristic value of current moment with inflection point is compared, and judges that the current time is
It is no to reach inflection point.
9. a kind of training sample generating device based on user characteristics knee value characterized by comprising
Memory, for storing computer executable program；
Data processing equipment requires to appoint in 1-4 for reading the computer executable program in the memory with perform claim
Based on the training sample generation method of user characteristics knee value described in one.
10. a kind of computer-readable medium, for storing computer-readable program, which is characterized in that the computer-readable journey
Sequence requires the training sample generation method based on user characteristics knee value described in any one of 1-4 for perform claim.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910330919.XA CN110147938A (en) | 2019-04-23 | 2019-04-23 | A kind of training sample generation method, device, system and recording medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910330919.XA CN110147938A (en) | 2019-04-23 | 2019-04-23 | A kind of training sample generation method, device, system and recording medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110147938A true CN110147938A (en) | 2019-08-20 |
Family
ID=67594351
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910330919.XA Pending CN110147938A (en) | 2019-04-23 | 2019-04-23 | A kind of training sample generation method, device, system and recording medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110147938A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111709832A (en) * | 2020-06-16 | 2020-09-25 | 中国银行股份有限公司 | Blacklist setting method and device, readable medium and equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107222328A (en) * | 2017-03-23 | 2017-09-29 | 广州华多网络科技有限公司 | Predict the method and device of saturation flex point |
| CN108460673A (en) * | 2017-02-22 | 2018-08-28 | 阿里巴巴集团控股有限公司 | A kind of processing method and processing device of training data |
| CN109191282A (en) * | 2018-08-23 | 2019-01-11 | 北京玖富普惠信息技术有限公司 | Methods of marking and system are monitored in a kind of loan of Behavior-based control model |
| CN109636591A (en) * | 2018-12-28 | 2019-04-16 | 浙江工业大学 | A kind of credit scoring card development approach based on machine learning |
-
2019
- 2019-04-23 CN CN201910330919.XA patent/CN110147938A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108460673A (en) * | 2017-02-22 | 2018-08-28 | 阿里巴巴集团控股有限公司 | A kind of processing method and processing device of training data |
| CN107222328A (en) * | 2017-03-23 | 2017-09-29 | 广州华多网络科技有限公司 | Predict the method and device of saturation flex point |
| CN109191282A (en) * | 2018-08-23 | 2019-01-11 | 北京玖富普惠信息技术有限公司 | Methods of marking and system are monitored in a kind of loan of Behavior-based control model |
| CN109636591A (en) * | 2018-12-28 | 2019-04-16 | 浙江工业大学 | A kind of credit scoring card development approach based on machine learning |
Non-Patent Citations (1)
| Title |
|---|
| 杨静: ""信用评分卡的建立与应用"", 《中国优秀硕士学位论文全文数据库 经济与管理科学辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111709832A (en) * | 2020-06-16 | 2020-09-25 | 中国银行股份有限公司 | Blacklist setting method and device, readable medium and equipment |
| CN111709832B (en) * | 2020-06-16 | 2023-07-14 | 中国银行股份有限公司 | Blacklist setting method, device, readable medium and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108846520A (en) | Overdue loan prediction technique, device and computer readable storage medium | |
| US7783551B1 (en) | Method and system for simulating risk factors in parametric models using risk neutral historical bootstrapping | |
| CN104616198A (en) | P2P (peer-to-peer) network lending risk prediction system based on text analysis | |
| Igamo et al. | The impact of electronic money on the efficiency of the payment system and the substitution of cash in Indonesia | |
| CN108446291A (en) | The real-time methods of marking and points-scoring system of user credit | |
| CN112541817A (en) | Marketing response processing method and system for potential customers of personal consumption loan | |
| CN109685644A (en) | A kind of customers' credit methods of marking and device based on transfer learning | |
| CN106651573A (en) | Business data processing method and apparatus | |
| CN107622444A (en) | Business access method and device, electronic equipment | |
| CN110009417A (en) | Target customer's screening technique, device, equipment and computer readable storage medium | |
| CN110533521A (en) | Method for early warning, device, equipment and readable storage medium storing program for executing after dynamic is borrowed | |
| CN109740129A (en) | Report form generation method, device, equipment and readable storage medium storing program for executing based on block chain | |
| CN110827143A (en) | Method, device and equipment for training credit scoring model | |
| WO2000054186A1 (en) | Financial forecasting system and method for risk assessment and management | |
| Mickiewicz et al. | Financial constraints in investment. Panel data results from Estonia, 1995-1999 | |
| CN110147938A (en) | A kind of training sample generation method, device, system and recording medium | |
| Engle | The risk that risk will change | |
| CN108629680A (en) | A kind of Risk Identification Method and system | |
| CN114493822A (en) | User default prediction pricing method and system based on transfer learning | |
| CN106874286A (en) | A kind of method and device for screening user characteristics | |
| CN110135972A (en) | A kind of method, apparatus, system and recording medium for improving user and moving branch rate | |
| CN117252677A (en) | Credit line determination method and device, electronic equipment and storage medium | |
| CN114328668A (en) | Method and device for generating deposit risk control strategy, terminal and storage medium | |
| CN111488463B (en) | Test corpus generation method and device and electronic equipment | |
| CN114529053A (en) | Neural network model construction method and prediction system for overdue amount prediction |
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 |