CN112862585A - Personal loan type bad asset risk rating method based on LightGBM decision tree algorithm - Google Patents
Personal loan type bad asset risk rating method based on LightGBM decision tree algorithm Download PDFInfo
- Publication number
- CN112862585A CN112862585A CN202011399404.4A CN202011399404A CN112862585A CN 112862585 A CN112862585 A CN 112862585A CN 202011399404 A CN202011399404 A CN 202011399404A CN 112862585 A CN112862585 A CN 112862585A
- Authority
- CN
- China
- Prior art keywords
- lightgbm
- data
- decision tree
- decision
- risk
- 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
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
- 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
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Abstract
The application discloses a personal loan type bad asset risk rating method based on a LightGBM decision tree algorithm, which comprises the following steps: collecting and processing data for risk rating; inputting the collected and processed data into a decision system constructed based on a LightGBM decision tree, and analyzing by the decision system according to the input data and a model constructed in the decision system; and outputting a rating result according to the analysis of the decision-making system. The method has the advantage of providing the method for grading the risk of the bad assets in the credit category based on the LightGBM decision tree algorithm, wherein the grading is carried out from the aspect of asset repayment possibility instead of the value of the assets.
Description
Technical Field
The application relates to a risk rating method, in particular to a personal loan type bad asset risk rating method based on a LightGBM decision tree algorithm.
Background
The loan-type poor assets belong to financial poor assets, and refer to secondary, suspicious and loss-type loans held by banks, financial poor creditors bought or taken over by financial asset management companies, and poor creditors held by other non-bank financial institutions. The existing credit-type bad asset rating mainly surrounds the value evaluation of the asset, the follow-up action is often asset disposal in an ABS asset package transfer mode, the risk rating in the aspect of the repayment possibility of the bad asset is almost not available, and the evaluation method is mostly based on personal experience, is very subjective and considers incomplete factors.
Disclosure of Invention
In order to solve the defects of the prior art, the application provides a method for grading the risk of the bad assets in the category of credit based on the LightGBM decision tree algorithm, which comprises the following steps: collecting and processing data for risk rating; inputting the collected and processed data into a decision system constructed based on a LightGBM decision tree, and analyzing by the decision system according to the input data and a model constructed in the decision system; and outputting a rating result according to the analysis of the decision making system.
Further, the method for grading the risk of the personal lending type poor assets based on the LightGBM decision tree algorithm further comprises the following steps: and constructing a model for constructing a decision system based on the LightGBM decision tree.
Further, the building of the model system for building the decision system based on the LightGBM decision tree includes the following steps: data extraction: three-party data based on API docking, such as Unionpay data, people's bank credit data, public accumulation fund/social security data, operator data, e-commerce transaction data, data of industrial and commercial companies, courts, information and the like obtained by compliance crawling of public websites, and user data of bad assets, and performing data matching; marking data: and marking samples according to the final refund condition of the historical bad assets fed back by the asset management company, wherein the performance period is given for three months, the refund in three months is a good user, and otherwise, the refund is a bad user.
Further, the sample enhancement: if there are fewer samples, then the samples need to be upsampled, i.e., sample enhanced. Multiple slice data may be generated through a sliding window in time series, with different slice data being equivalent to copying or enhancing the original single sample.
Further, the building of the model system for building the decision system based on the LightGBM decision tree includes the following steps: missing value filling: the deletion ratio is higher than 80 percent, and a direct deletion mode is adopted; if the ratio is lower than the ratio, missing data filling is carried out by adopting an interpolation method; data set partitioning: the data set is divided into a training set, a testing set and a verification set, and the sample ratio is 7: 2: 1.
further, the building of the model system for building the decision system based on the LightGBM decision tree includes the following steps: the method comprises the following steps: constructing features based on one or more combing basic indexes in user income, user liability, user credit, user external guarantee condition, user consumption and user whereabouts by combining time characteristics and common statistics; and (3) feature screening: screening is performed based on the importance and relevance of the features.
Further, the building of the model system for building the decision system based on the LightGBM decision tree includes the following steps: constructing a model: the model is built by adopting a LightGBM-LR framework, wherein the LightGBM is used for mapping derived features to high dimensionality; finally, the features of the high-dimensional vector space are used as the input of an LR model to predict the sample refund probability.
Further, the building of the model system for building the decision system based on the LightGBM decision tree includes the following steps: and (3) model evaluation: and respectively applying the constructed model to the test set sample and the verification set sample to respectively obtain the fund withdrawal probability predicted value of the test set sample and the fund withdrawal probability predicted value of the verification set sample.
Further, the model evaluation further comprises: and calculating the AUC value and the K-S value of the test sample set, and if the AUC value and the K-S value per se meet preset conditions and the difference value of the AUC value and the K-S value between a plurality of side-looking sample sets is within a preset difference value range, determining that the model is valid.
Further, the decision system outputs a money withdrawal probability value and outputs an evaluation grade according to the money withdrawal probability value.
The application has the advantages that: a method for grading the risk of the individual lending type bad assets based on a LightGBM decision tree algorithm is provided, wherein the grading is carried out from the perspective of asset repayment possibility, but not from the aspect of asset value.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, serve to provide a further understanding of the application and to enable other features, objects, and advantages of the application to be more apparent. The drawings and their description illustrate the embodiments of the invention and do not limit it. In the drawings:
fig. 1 is a schematic diagram illustrating steps of a method for grading the risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to an embodiment of the present application;
fig. 2 is a schematic diagram of a LightGBM-LR model in a method for grading a risk of an individual lending type bad asset based on a LightGBM decision tree algorithm according to an embodiment of the present application.
Detailed Description
In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only partial embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It should be understood that the data so used may be interchanged under appropriate circumstances such that embodiments of the application described herein may be used. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In this application, the terms "upper", "lower", "left", "right", "front", "rear", "top", "bottom", "inner", "outer", "middle", "vertical", "horizontal", "lateral", "longitudinal", and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings. These terms are used primarily to better describe the present application and its embodiments, and are not used to limit the indicated devices, elements or components to a particular orientation or to be constructed and operated in a particular orientation.
Moreover, some of the above terms may be used to indicate other meanings besides the orientation or positional relationship, for example, the term "on" may also be used to indicate some kind of attachment or connection relationship in some cases. The specific meaning of these terms in this application will be understood by those of ordinary skill in the art as appropriate.
Furthermore, the terms "mounted," "disposed," "provided," "connected," and "sleeved" are to be construed broadly. For example, it may be a fixed connection, a removable connection, or a unitary construction; can be a mechanical connection, or an electrical connection; may be directly connected, or indirectly connected through intervening media, or may be in internal communication between two devices, elements or components. The specific meaning of the above terms in the present application can be understood by those of ordinary skill in the art as appropriate.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
Referring to fig. 1 and 2, the present application provides a method for rating the risk of an individual lending type bad asset by using a LightGBM decision tree algorithm, which comprises the following steps: collecting and processing data for risk rating; inputting the collected and processed data into a decision system constructed based on a LightGBM decision tree, and analyzing by the decision system according to the input data and a model constructed in the decision system; and outputting a rating result according to the analysis of the decision-making system.
As a specific scheme, the method comprises the following steps:
the method comprises the following steps: and (6) data extraction. Three-party data based on API docking, such as Unionpay data, people's bank credit data, public accumulation fund/social security data, operator data, e-commerce transaction data, data of industry and commerce, court, information and the like obtained by compliance crawling of public websites, and user data of bad assets, and data matching is carried out.
Step two: and marking data. And marking samples according to the final refund condition of the historical bad assets fed back by the asset management company, wherein the performance period is given for three months, the refund in three months is a good user, and otherwise, the refund is a bad user.
Step three: and (4) sample enhancement. If there are fewer samples, then the samples need to be upsampled, i.e., sample enhanced. Multiple slice data may be generated through a sliding window in time series, with different slice data being equivalent to copying or enhancing the original single sample.
Step four: and filling missing values. The deletion ratio is higher than 80 percent, and a direct deletion mode is adopted; below this ratio, interpolation is used for missing data padding, and commonly used interpolation methods include averaging, mode, and the like. Missing values have important meanings themselves, and can be replaced by a discrete value as a class alone.
Step five: and (4) dividing the data set. The data set is divided into a training set, a testing set and a verification set, and the sample ratio is 7: 2: 1
Step six: and (5) feature construction. Based on the experience of business experts, the main dimensions influencing the refund of the bad assets are as follows: user income, user liability, user credit, user external guarantee, user consumption, user whereabouts, etc. According to the dimensions, the basic indexes are combed, and feature derivation is performed based on statistics such as mean, variance and ratio in combination with time characteristics. And constructing statistics such as the average value of the previous month of the current day, the average value of the next month of the current day, the ratio, the difference value, the trend and the like of the average value of the previous month and the average value of the next month of the current day, and calculating 8000+ dimensional characteristics.
Step seven: and (4) feature screening. Feature screening includes feature importance based screening and feature correlation based screening.
1) And (5) screening based on feature importance. From the data perspective, the higher the importance of a feature, the more valid information the feature contains, and the more important it is for the prediction of the result. The number of features should not be too large for generalization capability of the model. The model therefore screens the features according to their order of importance.
The training set data for these features is initially trained using LightGBM to obtain the importance of all features. Features were screened for importance greater than 0 as shown in FIG. 2.
2) And screening based on the characteristic relevance. Since too high correlation between features affects the stability of the model, further screening is required for the features with too high correlation.
And (3) calculating a characteristic correlation coefficient matrix, and selecting one of the characteristics with the absolute value of the correlation coefficient larger than 0.5 and similar business logic for retention (the gray marked in the figure 1 is the characteristic deleted according to the correlation).
Step eight: and (5) constructing a model. After feature screening is completed, the model is built by adopting the LightGBM-LR framework. Wherein the LightGBM is used to map derived features to high dimensionality; finally, the features of the high-dimensional vector space are used as the input of an LR model to predict the sample refund probability.
The model construction process is as follows:
1) constructing a LightGBM model of 10 decision trees with 128 leaf nodes per tree to map features to a vector space of 1280 dimensions (128 x 10);
2) inputting each log data as an x, converting the 43-dimensional features constructed in the early stage into 1280-dimensional vector space through a LightGBM model, and encoding the newly generated vector by one-hot;
3) and predicting the suspicion degree of the sample by using the 1280-dimensional vector space as the input of an LR model, and judging whether electricity is stolen or not. To prevent overfitting, the LR model adds L2 regularization.
The final model predicts the probability of refund of the bad asset
Step nine: and (5) evaluating and optimizing the model. And based on the model parameters predicted in the step eight, respectively applying the model parameters to the test set samples and the verification set samples to respectively obtain the refund probability predicted values of the test set samples and the refund probability predicted values of the verification set samples. Finally calculating the AUC value and the K-S value of the three sample sets, if KS > is 0.3 and AUC >0.7, and the AUC and KS of the three sample sets are not different, which indicates that the model is stable and effective and can be applied; otherwise, optimization is needed, and the step eight is repeated to adjust parameters.
Step ten: and (4) grading the bad assets, and carrying out grading mapping on the assets with the probability of being greater than 0.5 based on the refund probability value predicted by the model, wherein the grades are 0.8-0.9 for A grade 0.8-0.9, 0.7-0.8 for B grade 0.6-0.7 for C grade 0.5-0.6 for E grade 0.5 or less for F grade 0.9-1.0.
The method for grading the risk of the poor assets is based on big data, adopts the LightGBM algorithm, does not need to depend on subjective experience of people, and finally outputs the risk grading result of the poor assets. In addition, solution 1. the bad asset rating is rated from the asset refund likelihood perspective rather than the value of the asset itself. The method for grading the bad assets adopts a big data mode, the LightGBM is adopted as the algorithm, and the algorithm has the advantages of high efficiency and rapidness. The model generates 10 decision trees, 128 leaf nodes and 1280-dimensional vector space, and information is fully utilized. And converting the final rating result according to the money return probability, and rating the final rating result into 7 grades such as ABCDEF and the like.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (10)
1. A personal loan type bad asset risk rating method based on a LightGBM decision tree algorithm is characterized in that:
the method for grading the risk of the individual lending type poor assets based on the LightGBM decision tree algorithm comprises the following steps:
collecting and processing data for risk rating;
inputting the collected and processed data into a decision system constructed based on a LightGBM decision tree, and analyzing by the decision system according to the input data and a model constructed in the decision system;
and outputting a rating result according to the analysis of the decision making system.
2. The method of grading a risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to claim 1, wherein:
the method for grading the risk of the individual lending type poor assets based on the LightGBM decision tree algorithm further comprises the following steps:
and constructing a model for constructing a decision system based on the LightGBM decision tree.
3. The method of grading a risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to claim 1, wherein:
the method for constructing the model system for constructing the decision system based on the LightGBM decision tree comprises the following steps:
data extraction: three-party data based on API docking, such as Unionpay data, people's bank credit data, public accumulation fund/social security data, operator data, e-commerce transaction data, data of industrial and commercial companies, courts, information and the like obtained by compliance crawling of public websites, and user data of bad assets, and performing data matching;
marking data: and marking samples according to the final refund condition of the historical bad assets fed back by the asset management company, wherein the performance period is given for three months, the refund in three months is a good user, and otherwise, the refund is a bad user.
4. The method of grading a risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to claim 1, wherein:
sample enhancement: if there are fewer samples, then the samples need to be upsampled, i.e., sample enhanced. Multiple slice data may be generated through a sliding window in time series, with different slice data being equivalent to copying or enhancing the original single sample.
5. The method of grading a risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to claim 1, wherein:
the method for constructing the model system for constructing the decision system based on the LightGBM decision tree comprises the following steps:
missing value filling: the deletion ratio is higher than 80 percent, and a direct deletion mode is adopted; if the ratio is lower than the ratio, missing data filling is carried out by adopting an interpolation method;
data set partitioning: the data set is divided into a training set, a testing set and a verification set, and the sample ratio is 7: 2: 1.
6. the method of grading a risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to claim 1, wherein:
the method for constructing the model system for constructing the decision system based on the LightGBM decision tree comprises the following steps:
the method comprises the following steps: constructing features based on one or more combing basic indexes in user income, user liability, user credit, user external guarantee condition, user consumption and user whereabouts by combining time characteristics and common statistics;
and (3) feature screening: screening is performed based on the importance and relevance of the features.
7. The method of grading a risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to claim 1, wherein:
the method for constructing the model system for constructing the decision system based on the LightGBM decision tree comprises the following steps:
constructing a model: the model is built by adopting a LightGBM-LR framework, wherein the LightGBM is used for mapping derived features to high dimensionality; finally, the features of the high-dimensional vector space are used as the input of an LR model to predict the sample refund probability.
8. The method of grading a risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to claim 1, wherein:
the method for constructing the model system for constructing the decision system based on the LightGBM decision tree comprises the following steps:
and (3) model evaluation: and respectively applying the constructed model to the test set sample and the verification set sample to respectively obtain the fund withdrawal probability predicted value of the test set sample and the fund withdrawal probability predicted value of the verification set sample.
9. The method of grading a risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to claim 1, wherein:
the model evaluation further comprises:
and calculating the AUC value and the K-S value of the test sample set, and if the AUC value and the K-S value per se meet preset conditions and the difference value of the AUC value and the K-S value between a plurality of side-looking sample sets is within a preset difference value range, determining that the model is valid.
10. The method of grading a risk of an individual lending type poor asset based on a LightGBM decision tree algorithm according to claim 1, wherein:
and the decision system outputs a money return probability value and outputs an evaluation grade according to the money return probability value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011399404.4A CN112862585A (en) | 2020-12-02 | 2020-12-02 | Personal loan type bad asset risk rating method based on LightGBM decision tree algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011399404.4A CN112862585A (en) | 2020-12-02 | 2020-12-02 | Personal loan type bad asset risk rating method based on LightGBM decision tree algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112862585A true CN112862585A (en) | 2021-05-28 |
Family
ID=75996878
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011399404.4A Pending CN112862585A (en) | 2020-12-02 | 2020-12-02 | Personal loan type bad asset risk rating method based on LightGBM decision tree algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112862585A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113610628A (en) * | 2021-07-28 | 2021-11-05 | 浙江惠瀜网络科技有限公司 | Pre-loan risk event investigation method and device and electronic equipment |
| CN114154866A (en) * | 2021-12-02 | 2022-03-08 | 北京顶象技术有限公司 | Marketing enterprise financial risk early warning method and system |
| CN115660834A (en) * | 2022-12-23 | 2023-01-31 | 河北雄安舜耕数据科技有限公司 | Individual loan risk assessment method based on decision tree |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106952155A (en) * | 2017-03-08 | 2017-07-14 | 深圳前海纵腾金融科技服务有限公司 | A kind of collection method and device based on credit scoring |
| CN110009479A (en) * | 2019-03-01 | 2019-07-12 | 百融金融信息服务股份有限公司 | Credit assessment method and device, storage medium, computer equipment |
| CN111311401A (en) * | 2020-03-30 | 2020-06-19 | 百维金科(上海)信息科技有限公司 | Financial default probability prediction model based on LightGBM |
| CN111652710A (en) * | 2020-06-03 | 2020-09-11 | 北京化工大学 | Personal credit risk assessment method based on ensemble tree feature extraction and Logistic regression |
-
2020
- 2020-12-02 CN CN202011399404.4A patent/CN112862585A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106952155A (en) * | 2017-03-08 | 2017-07-14 | 深圳前海纵腾金融科技服务有限公司 | A kind of collection method and device based on credit scoring |
| CN110009479A (en) * | 2019-03-01 | 2019-07-12 | 百融金融信息服务股份有限公司 | Credit assessment method and device, storage medium, computer equipment |
| CN111311401A (en) * | 2020-03-30 | 2020-06-19 | 百维金科(上海)信息科技有限公司 | Financial default probability prediction model based on LightGBM |
| CN111652710A (en) * | 2020-06-03 | 2020-09-11 | 北京化工大学 | Personal credit risk assessment method based on ensemble tree feature extraction and Logistic regression |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113610628A (en) * | 2021-07-28 | 2021-11-05 | 浙江惠瀜网络科技有限公司 | Pre-loan risk event investigation method and device and electronic equipment |
| CN114154866A (en) * | 2021-12-02 | 2022-03-08 | 北京顶象技术有限公司 | Marketing enterprise financial risk early warning method and system |
| CN115660834A (en) * | 2022-12-23 | 2023-01-31 | 河北雄安舜耕数据科技有限公司 | Individual loan risk assessment method based on decision tree |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Albashrawi | Detecting financial fraud using data mining techniques: A decade review from 2004 to 2015 | |
| CN110334737B (en) | Customer risk index screening method and system based on random forest | |
| Yue et al. | A review of data mining-based financial fraud detection research | |
| CN112862585A (en) | Personal loan type bad asset risk rating method based on LightGBM decision tree algorithm | |
| CN106897918A (en) | A kind of hybrid machine learning credit scoring model construction method | |
| KR102009309B1 (en) | Management automation system for financial products and management automation method using the same | |
| CN110909984B (en) | Business data processing model training method, business data processing method and device | |
| CN105308640A (en) | Methods and systems for automatically generating high quality adverse action notifications | |
| Yotsawat et al. | A novel method for credit scoring based on cost-sensitive neural network ensemble | |
| US11263644B2 (en) | Systems and methods for detecting unauthorized or suspicious financial activity | |
| CN112700324A (en) | User loan default prediction method based on combination of Catboost and restricted Boltzmann machine | |
| CN115545886A (en) | Overdue risk identification method, overdue risk identification device, overdue risk identification equipment and storage medium | |
| Chen | A hybrid knowledge-sharing model for corporate foreign investment in China’s construction market | |
| CN110689437A (en) | Communication construction project financial risk prediction method based on random forest | |
| Khder et al. | The impact of implementing data mining in business intelligence | |
| CN111062806A (en) | Personal finance credit risk evaluation method, system and storage medium | |
| Balcı et al. | Hierarchies in communities of UK stock market from the perspective of Brexit | |
| Hong et al. | Generating and Validating Synthetic Training Data for Predicting Bankruptcy of Individual Businesses. | |
| da Silva Mattos et al. | Bankruptcy prediction with low-quality financial information | |
| Xi et al. | Improved AHP model and neural network for consumer finance credit risk assessment | |
| CN114219606A (en) | Power data-based power consumption enterprise credit evaluation method and system | |
| CN113139868A (en) | Customer credit evaluation and loan potential analysis method based on electricity consumption big data | |
| CN113298393A (en) | Vehicle loan risk assessment method based on regression algorithm | |
| CN114154682A (en) | Customer loan yield grade prediction method and system | |
| Eko et al. | Improving credit scoring model of mortgage financing with smote methods in sharia banking |
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 |