Credit Risk Management of P2P Network Lending

Li, Dongmei; Na, Sanggyun; Ding, Tao; Liu, Congchong

doi:10.17559/tv-20200210110508

Cited by 5 publications

(2 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To measure and compare the performance of machine learning models in meta path features, six commonly used machine learning classification algorithms, namely SVM, NN, K-NN, RF, AdaBoost, and GBDT, are used for comparative analysis. The result is compared with the most popular credit scoring model, logistic regression, and an upgraded credit risk measuring model (G-LR) based on logistic regression, which is proposed in [34]. Each set of experiments was performed 50 times, and the average of each result was taken as the final result.…”

Section: Results Analysismentioning

confidence: 99%

Modeling Selection for Credit Risk Measurement: Based on Meta Path Features

2023

Teh. vjesn.

View full text Add to dashboard Cite

As the application scope of machine learning expands, studies of credit risk measurement have also witnessed extensive development. An increasing number of studies showed that models based on machine learning algorithms could be used as a substantive solution for credit risk modeling. Recently, path-based features showed their advantages in risk measurement of the rich semantic and relational information it contains. However, studies have yet to be probed into the field that combines meta path features and machine learning models to measure the credit risk of the enterprise. In response to this problem, we compare the performance of machine learning models in terms of meta path features to find suitable machine learning models for meta path features. This paper compares six commonly used machine learning classification models, including neural networks, support vector machine, k-nearest neighbor, random forest, AdaBoost, and GBDT. Experiments on three listed small and medium-sized enterprises datasets in China. We found that neural networks, support vector machines, and the GBDT model perform better than other machine learning models. These are potential classifiers for small and medium-sized enterprises' credit risk measurement.

show abstract

Section: Results Analysismentioning

confidence: 99%

Modeling Selection for Credit Risk Measurement: Based on Meta Path Features

2023

Teh. vjesn.

View full text Add to dashboard Cite

show abstract

“…However, P2P lending has a high failure rate for borrowers to repay their loans and P2P lending also has low-interest rates, which is known as credit risk [1]- [3]. Credit risk is a major problem in P2P Lending because payers who fail to make loan payments make the lender suffer losses [2]. To minimize credit risk, a system is needed that can assist in determining credit risk.…”

Section: Introductionmentioning

confidence: 99%

Credit Risk Detection in Peer-to-Peer Lending Using CatBoost

Nasution,

Saadah,

Yunanto

2023

J. RESTI (Rekayasa Sist. Teknol. Inf.)

View full text Add to dashboard Cite

P2P lending (Peer-to-peer lending) is widely used by private borrowers, small businesses, and MSMEs because P2P lending allows individuals and businesses to be able to lend money directly from lenders without the stringent requirements and criteria of traditional banks and financial institutions. However, P2P lending has a credit risk problem characterized by a high failure rate for borrowers to repay their loans. Therefore, a system was necessary to detect credit risk for minimizing P2P lending credit risk. In this study, a system had been built using the CatBoost method, which the used dataset was taken from loan dataset of the Bondora company. To measure the performance of the CatBoost algorithm, an evaluation matrix used ROC (Receiver Operating Characteristics) curves and AUC (Area Under Curve) had been conducted. The experiment consists of three scenarios which the best result regard to scenario 2 with a data splitting of 90:10. It was caused by the result of AUC value’s 0.80329 compared to scenario 1 with a data splitting of 80:20 with the AUC value around 0.789583, and scenario 3 with a data splitting of 70:30 with the AUC value around 0.781066, respectively.

show abstract