CN111242441A - Adaptive parameter fitting method suitable for small and micro enterprise risk control model - Google Patents
Adaptive parameter fitting method suitable for small and micro enterprise risk control model Download PDFInfo
- Publication number
- CN111242441A CN111242441A CN202010009481.8A CN202010009481A CN111242441A CN 111242441 A CN111242441 A CN 111242441A CN 202010009481 A CN202010009481 A CN 202010009481A CN 111242441 A CN111242441 A CN 111242441A
- Authority
- CN
- China
- Prior art keywords
- coefficients
- subclass
- subclasses
- class
- significance probability
- 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.)
- Granted
Links
Images
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
- 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
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Operations Research (AREA)
- Evolutionary Biology (AREA)
- Mathematical Optimization (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Marketing (AREA)
- Probability & Statistics with Applications (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Algebra (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Complex Calculations (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to a self-adaptive parameter fitting method suitable for a small micro-enterprise risk control model. A self-adaptive parameter fitting method suitable for a small and micro enterprise risk control model is disclosed. Compared with the prior art, the LR method for training the model can effectively prevent the model from being over-fitted and enable the model to be rapidly converged, and finally the model obtains the coefficient weights of various classes and subclasses, so that the model can be flexibly split and combined for application, and can rapidly and effectively deal with the sudden situation caused by the deficiency of a certain subclass such as a certain data source.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a self-adaptive parameter fitting method suitable for a small micro-enterprise risk control model.
Background
In financial risk control, data modeling is an increasingly important wind control technique. In a traditional data model facing a consumer end, because the sample size is large, the risk coefficient of each feature can be automatically fitted according to a certain parameter estimation method, and a good prediction effect can be obtained. However, in a data model facing an enterprise side, especially a small-scale enterprise side, the sample size is often much smaller, and the accuracy of the automatically fitted coefficients is not good, so that it is difficult to effectively measure the relative importance of each feature.
Disclosure of Invention
The LR method is used for training the model, so that overfitting of the model can be effectively prevented, the model can be quickly converged, coefficient weights of various classes and subclasses can be obtained by the model finally, the model can be flexibly split and combined for application, and sudden situations caused by the fact that a certain subclass such as a certain data source is lost can be quickly and effectively dealt with.
In order to achieve the purpose, a self-adaptive parameter fitting method suitable for a small enterprise risk control model is designed, and is characterized in that: the specific method comprises the following steps:
(1) collecting a sample;
(2) classifying the collected sample characteristics, and classifying the characteristics into a large class A (namely a risk module) according to the characteristics with the same or similar property data;
(3) each large class is from some data sources and is classified as a subclass b according to the characteristics of the same or similar data sources;
(4) after all the data are finished, each large class comprises a plurality of subclasses, and each subclass comprises a plurality of characteristics;
(5) counting the discrimination of each characteristic on good and bad samples, reserving the characteristic with obvious discrimination, and rejecting the characteristic without discrimination or with unobvious discrimination;
(6) scoring each reserved feature, mapping the score to a corresponding risk interval according to the discrimination of each feature on the quality level, and enabling the score value range of each feature to be [0,100 ];
(7) after the grading of each feature is finished, grading each subclass of the upper-level attribute of the feature; mapping the score range of each subclass to [0,100 ];
(8) establishing a logistic regression model, and calculating an estimated value and a significance probability value of the subclasses by adopting a maximum likelihood estimation method;
(9) judging whether the significance probability value of the subclass is less than 0.05, if so, summing the products of the scores of the subclass and the corresponding coefficients to obtain an initial score of the subclass; otherwise, returning to the step (3);
(10) establishing a logistic regression model, and calculating a large-class estimation value and a significance probability value by adopting a maximum likelihood estimation method;
(11) judging whether the significance probability value of the large class is less than 0.05, if so, summing the products of the scores of the large class and the corresponding coefficients to obtain a total initial score; otherwise, returning to the step (3).
In the step (7), scoring the upper class of the subclasses according to the score of each subclass, the specific method is as follows:
(71) assuming that m features x1, x2, x3, …, xm and 1 binary variable y are contained in a sample with the size of N, the samples are classified into i major classes A1, A2, … and Ai after the processing from the step (1) to the step (4), and each class contains N subclasses bi1, bi2, … and bin;
(72) respectively establishing a logistic regression model for the subclasses in each major class Ai: ln p/(1-p) = ri0 + ri1bi 1+ ri2bi 2+. + rinbin, where p is the probability of y =1, using maximum likelihood estimation, the estimated values of coefficients ri0, ri1, …, rin and significance probability values pi0, pi1, …, pin are calculated;
(73) observing whether the significance probability value of each coefficient is less than 0.05, if the significance probability value of a certain coefficient is greater than or equal to 0.05, returning to the steps (3) to (4) to perform feature and subclass scoring again until all the coefficients are significant;
(74) summing the products of the scores of the subclasses bi1, bi2 and … and the corresponding coefficients by the aid of bi1 ri1+ bi2 ri2+ … + bin rin to obtain initial scores of the major classes Ai, mapping the maximum value of the initial scores of the major classes Ai to 100, and performing corresponding proportional scaling on the coefficients to obtain final coefficients (weights) of all the subclasses in the major classes Ai;
(75) accordingly, a logistic regression model is established for each large class Ai: ln P/(1-P) = R0 + R1a 1+ R2a2+. + RiAi, where P is the probability of y =1, and the estimated values of coefficients R0, R1, …, Ri and significance probability values P0, P1, …, Pi are calculated by the maximum likelihood estimation method;
(76) observing whether the significance probability value of each coefficient is less than 0.05, if the significance probability value of a certain coefficient is greater than or equal to 0.05, returning to the step (74) to the step (75), and carrying out class scoring again until all the coefficients are significant;
(77) and summing the products of the scores of the large classes A1, A2, … and Ai and the corresponding coefficients A1R 1+ A2R 2+. + AiRi to obtain a total initial score, mapping the maximum value of the total initial score to 100, and performing corresponding scaling on the coefficients to obtain the final coefficients (weights) of the classes Ai.
Step (5), counting the quality samples of each characteristic pair: good and bad samples are generally distinguished by the number of overdue samples, which are bad samples after two or more months, and good samples without overdue samples.
Compared with the prior art, the LR method training model can effectively prevent the model from being over-fitted and enable the model to be rapidly converged, and finally the model obtains the coefficient weights of each major class and subclass, so that the model can be flexibly split and combined for application, and can rapidly and effectively deal with the sudden situation caused by the deficiency of a subclass such as a data source.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The invention is further illustrated below with reference to the accompanying drawings.
As shown in fig. 1, a method for fitting adaptive parameters suitable for a risk control model of a small micro enterprise includes the following specific steps:
(1) collecting a sample;
(2) classifying the collected sample characteristics into a large class A (namely a risk module) according to characteristics with the same or similar property data, such as 'A1 is a multi-head credit class' and 'A2 is a historical performance class' in FIG. 1;
(3) after classification, each large class comprises a plurality of characteristics with the same or similar properties, each large class is from some data sources and is classified as a subclass b according to the characteristics of the same or similar data sources;
(4) after all the data are finished, each large class comprises a plurality of subclasses, and each subclass comprises a plurality of characteristics; for example, the following subclasses of A1 in FIG. 1 can be classified as "b 11 network audit", "b 12 short message audit", etc.;
(5) counting the discrimination of each characteristic on good and bad samples, reserving the characteristic with obvious discrimination, and rejecting the characteristic without discrimination or with unobvious discrimination; culling features "12 month query times" as in fig. 1;
(6) scoring each reserved feature, mapping the score to a corresponding risk interval according to the discrimination of each feature on the quality level, and enabling the score value range of each feature to be [0,100 ];
(7) after the grading of each feature is finished, grading each subclass of the upper-level attribute of the feature; mapping the score range of each subclass to [0,100 ];
(8) establishing a logistic regression model, and calculating an estimated value and a significance probability value of the subclasses by adopting a maximum likelihood estimation method;
(9) judging whether the significance probability value of the subclass is less than 0.05, if so, summing the products of the scores of the subclass and the corresponding coefficients to obtain an initial score of the subclass; otherwise, returning to the step (3);
(10) establishing a logistic regression model, and calculating a large-class estimation value and a significance probability value by adopting a maximum likelihood estimation method;
(11) judging whether the significance probability value of the large class is less than 0.05, if so, summing the products of the scores of the large class and the corresponding coefficients to obtain a total initial score; otherwise, returning to the step (3).
In the step (7), scoring the upper class of the subclasses according to the score of each subclass, the specific method is as follows:
(71) assuming that m features x1, x2, x3, …, xm and 1 binary variable y are contained in a sample with the size of N, the samples are classified into i major classes A1, A2, … and Ai after the processing from the step (1) to the step (4), and each class contains N subclasses bi1, bi2, … and bin;
(72) respectively establishing a logistic regression model for the subclasses in each major class Ai: ln p/(1-p) = ri0 + ri1bi 1+ ri2bi 2+. + rinbin, where p is the probability of y =1, using maximum likelihood estimation, the estimated values of coefficients ri0, ri1, …, rin and significance probability values pi0, pi1, …, pin are calculated;
(73) observing whether the significance probability value of each coefficient is less than 0.05, if the significance probability value of a certain coefficient is greater than or equal to 0.05, returning to the steps (3) to (4) to perform feature and subclass scoring again until all the coefficients are significant;
(74) summing the products of the scores of the subclasses bi1, bi2 and … and the corresponding coefficients by the aid of bi1 ri1+ bi2 ri2+ … + bin rin to obtain initial scores of the major classes Ai, mapping the maximum value of the initial scores of the major classes Ai to 100, and performing corresponding proportional scaling on the coefficients to obtain final coefficients (weights) of all the subclasses in the major classes Ai;
(75) accordingly, a logistic regression model is established for each large class Ai: ln P/(1-P) = R0 + R1a 1+ R2a2+. + RiAi, where P is the probability of y =1, and the estimated values of coefficients R0, R1, …, Ri and significance probability values P0, P1, …, Pi are calculated by the maximum likelihood estimation method;
(76) observing whether the significance probability value of each coefficient is less than 0.05, if the significance probability value of a certain coefficient is greater than or equal to 0.05, returning to the step (74) to the step (75), and carrying out class scoring again until all the coefficients are significant;
(77) and summing the products of the scores of the large classes A1, A2, … and Ai and the corresponding coefficients A1R 1+ A2R 2+. + AiRi to obtain a total initial score, mapping the maximum value of the total initial score to 100, and performing corresponding scaling on the coefficients to obtain the final coefficients (weights) of the classes Ai.
Step (5), counting the quality samples of each characteristic pair: good and bad samples are generally distinguished by the number of overdue samples, which are bad samples after two or more months, and good samples without overdue samples.
The advantages of the invention are mainly embodied in the following 3 aspects:
1. the training data amount is small, the required data amount is hundreds of levels and above, and other methods usually require thousands of levels or even tens of thousands of levels and above.
2. The bottom layer characteristics are screened and normalized to [0,100], the construction process of subclass grading is flexible, and a maximum value method, a weighted average method and the like can be flexibly adopted by combining with expert business experience.
3. The model is trained by adopting an LR + LR method, so that overfitting of the model can be effectively prevented, the model can be quickly converged, coefficient weights of various major classes and subclasses can be obtained by the model finally, the model can be flexibly split and combined for application, and sudden situations caused by the deletion of a certain subclass such as a certain data source can be quickly and effectively dealt with.
Claims (3)
1. A self-adaptive parameter fitting method suitable for a small and micro enterprise risk control model is characterized by comprising the following steps: the specific method comprises the following steps:
(1) collecting a sample;
(2) classifying the collected sample characteristics, and classifying the characteristics into a large class A (namely a risk module) according to the characteristics with the same or similar property data;
(3) each large class is from some data sources and is classified as a subclass b according to the characteristics of the same or similar data sources;
(4) after all the data are finished, each large class comprises a plurality of subclasses, and each subclass comprises a plurality of characteristics;
(5) counting the discrimination of each characteristic on good and bad samples, reserving the characteristic with obvious discrimination, and rejecting the characteristic without discrimination or with unobvious discrimination;
(6) scoring each reserved feature, mapping the score to a corresponding risk interval according to the discrimination of each feature on the quality level, and enabling the score value range of each feature to be [0,100 ];
(7) after the grading of each feature is finished, grading each subclass of the upper-level attribute of the feature; mapping the score range of each subclass to [0,100 ];
(8) establishing a logistic regression model, and calculating an estimated value and a significance probability value of the subclasses by adopting a maximum likelihood estimation method;
(9) judging whether the significance probability value of the subclass is less than 0.05, if so, summing the products of the scores of the subclass and the corresponding coefficients to obtain an initial score of the subclass; otherwise, returning to the step (3);
(10) establishing a logistic regression model, and calculating a large-class estimation value and a significance probability value by adopting a maximum likelihood estimation method;
(11) judging whether the significance probability value of the large class is less than 0.05, if so, summing the products of the scores of the large class and the corresponding coefficients to obtain a total initial score; otherwise, returning to the step (3).
2. The adaptive parameter fitting method suitable for the risk control model of the small micro enterprise according to claim 1, wherein: in the step (7), scoring the upper class of the subclasses according to the score of each subclass, the specific method is as follows:
(71) assuming that m features x1, x2, x3, …, xm and 1 binary variable y in a sample with size N are classified into i major classes a1, a2, …, Ai after the processing of steps (1) to (4) in claim 1, each class containing N subclasses bi1, bi2, …, bin;
(72) respectively establishing a logistic regression model for the subclasses in each major class Ai: ln p/(1-p) = ri0 + ri1bi 1+ ri2bi 2+. + rinbin, where p is the probability of y =1, using maximum likelihood estimation, the estimated values of coefficients ri0, ri1, …, rin and significance probability values pi0, pi1, …, pin are calculated;
(73) observing whether the significance probability value of each coefficient is less than 0.05, if the significance probability value of a certain coefficient is greater than or equal to 0.05, returning to the steps (3) to (4) in the claim 1, and scoring the characteristics and subclasses again until all the coefficients are significant;
(74) summing the products of the scores of the subclasses bi1, bi2 and … and the corresponding coefficients by the aid of bi1 ri1+ bi2 ri2+ … + bin rin to obtain initial scores of the major classes Ai, mapping the maximum value of the initial scores of the major classes Ai to 100, and performing corresponding proportional scaling on the coefficients to obtain final coefficients (weights) of all the subclasses in the major classes Ai;
(75) accordingly, a logistic regression model is established for each large class Ai: ln P/(1-P) = R0 + R1a 1+ R2a2+. + RiAi, where P is the probability of y =1, and the estimated values of coefficients R0, R1, …, Ri and significance probability values P0, P1, …, Pi are calculated by the maximum likelihood estimation method;
(76) observing whether the significance probability value of each coefficient is less than 0.05, if the significance probability value of a certain coefficient is greater than or equal to 0.05, returning to the step (74) to the step (75), and carrying out class scoring again until all the coefficients are significant;
(77) and summing the products of the scores of the large classes A1, A2, … and Ai and the corresponding coefficients A1R 1+ A2R 2+. + AiRi to obtain a total initial score, mapping the maximum value of the total initial score to 100, and performing corresponding scaling on the coefficients to obtain the final coefficients (weights) of the classes Ai.
3. The adaptive parameter fitting method suitable for the risk control model of the small micro enterprise according to claim 1, wherein: step (5), counting the quality samples of each characteristic pair: good and bad samples are generally distinguished by the number of overdue samples, which are bad samples after two or more months, and good samples without overdue samples.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010009481.8A CN111242441B (en) | 2020-01-06 | 2020-01-06 | Self-adaptive parameter fitting method suitable for small micro-enterprise risk control model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010009481.8A CN111242441B (en) | 2020-01-06 | 2020-01-06 | Self-adaptive parameter fitting method suitable for small micro-enterprise risk control model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111242441A true CN111242441A (en) | 2020-06-05 |
| CN111242441B CN111242441B (en) | 2023-06-30 |
Family
ID=70879748
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010009481.8A Active CN111242441B (en) | 2020-01-06 | 2020-01-06 | Self-adaptive parameter fitting method suitable for small micro-enterprise risk control model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111242441B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050111367A1 (en) * | 2003-11-26 | 2005-05-26 | Hung-Hsiang Jonathan Chao | Distributed architecture for statistical overload control against distributed denial of service attacks |
| CN103942604A (en) * | 2013-01-18 | 2014-07-23 | 上海安迪泰信息技术有限公司 | Prediction method and system based on forest discrimination model |
| CN105930430A (en) * | 2016-04-19 | 2016-09-07 | 北京邮电大学 | Non-cumulative attribute based real-time fraud detection method and apparatus |
| US10078743B1 (en) * | 2013-01-31 | 2018-09-18 | Narus, Inc. | Cross identification of users in cyber space and physical world |
| CN108985361A (en) * | 2018-07-02 | 2018-12-11 | 北京金睛云华科技有限公司 | A kind of malicious traffic stream detection implementation method and device based on deep learning |
| CN109040143A (en) * | 2018-10-18 | 2018-12-18 | 中国联合网络通信集团有限公司 | A kind of detection method and device of BGP anomalous event |
| CN110378786A (en) * | 2019-07-29 | 2019-10-25 | 中国工商银行股份有限公司 | Model training method, promise breaking conduction Risk Identification Method, device and storage medium |
| CN110443304A (en) * | 2019-08-06 | 2019-11-12 | 民生科技有限责任公司 | A kind of business risk appraisal procedure based on machine learning model |
-
2020
- 2020-01-06 CN CN202010009481.8A patent/CN111242441B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050111367A1 (en) * | 2003-11-26 | 2005-05-26 | Hung-Hsiang Jonathan Chao | Distributed architecture for statistical overload control against distributed denial of service attacks |
| CN103942604A (en) * | 2013-01-18 | 2014-07-23 | 上海安迪泰信息技术有限公司 | Prediction method and system based on forest discrimination model |
| US10078743B1 (en) * | 2013-01-31 | 2018-09-18 | Narus, Inc. | Cross identification of users in cyber space and physical world |
| CN105930430A (en) * | 2016-04-19 | 2016-09-07 | 北京邮电大学 | Non-cumulative attribute based real-time fraud detection method and apparatus |
| CN108985361A (en) * | 2018-07-02 | 2018-12-11 | 北京金睛云华科技有限公司 | A kind of malicious traffic stream detection implementation method and device based on deep learning |
| CN109040143A (en) * | 2018-10-18 | 2018-12-18 | 中国联合网络通信集团有限公司 | A kind of detection method and device of BGP anomalous event |
| CN110378786A (en) * | 2019-07-29 | 2019-10-25 | 中国工商银行股份有限公司 | Model training method, promise breaking conduction Risk Identification Method, device and storage medium |
| CN110443304A (en) * | 2019-08-06 | 2019-11-12 | 民生科技有限责任公司 | A kind of business risk appraisal procedure based on machine learning model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111242441B (en) | 2023-06-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110009479B (en) | Credit evaluation method and device, storage medium and computer equipment | |
| CN111461180B (en) | Sample classification method, device, computer equipment and storage medium | |
| CN111461216B (en) | Case risk identification method based on machine learning | |
| CN108876034B (en) | Improved Lasso + RBF neural network combination prediction method | |
| CN106157132A (en) | Credit risk monitoring system and method | |
| US7805395B2 (en) | System and method for automated population splitting to assist task management through analytics | |
| CN110930250A (en) | Enterprise credit risk prediction method and system, storage medium and electronic equipment | |
| CN111796957A (en) | Transaction abnormal root cause analysis method and system based on application log | |
| CN113554310A (en) | Enterprise credit dynamic evaluation model based on intelligent contract | |
| CN113065969A (en) | Enterprise scoring model construction method, enterprise scoring method, medium and electronic device | |
| CN114584601B (en) | User loss identification and intervention method, system, terminal and medium | |
| CN117453764A (en) | Data mining analysis method | |
| CN116468536A (en) | Automatic risk control rule generation method | |
| CN109087146A (en) | The prediction technique and system of movie theatre box-office income | |
| Glennon et al. | Development and validation of credit scoring models | |
| CN111242441A (en) | Adaptive parameter fitting method suitable for small and micro enterprise risk control model | |
| CN112712270B (en) | Information processing method, device, equipment and storage medium | |
| KR100727555B1 (en) | Creating method for decision tree using time-weighted entropy and recording medium thereof | |
| CN113191569A (en) | Enterprise management method and system based on big data | |
| Gupta et al. | Forecasting Customer Churn in the Telecommunications Industry | |
| Xu et al. | Churn prediction in telecom using a hybrid two-phase feature selection method | |
| CN115270140B (en) | Software security vulnerability management method and system | |
| CN118429728B (en) | Chemical fertilizer particle detection method based on small sample learning | |
| Huang | Recognition Algorithm of Financial Data Validity of Listed Enterprises Based on Classification Regression Decision Tree | |
| CN118710388A (en) | Credit risk model construction method, system and storage medium based on reject inference |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 201900 room 502A, building 2, No. 439, Shitai Road, Baoshan District, Shanghai Applicant after: Shanghai Fuli Technology Co.,Ltd. Address before: 201900 room 502A, building 2, No. 439, Shitai Road, Baoshan District, Shanghai Applicant before: SHANGHAI FULI FINANCIAL INFORMATION SERVICE Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |