"PREDICTIVE MODELING FOR CREDIT SCORING AND  COMPARATIVE DATA ANALYSIS USING MACHINE  LEARNING METHODS FOR ETHIOPIAN BANK’S"

ESKINDER, ESHETU ABEBE

dc.contributor.author	ESKINDER, ESHETU ABEBE
dc.date.accessioned	2024-02-28T12:06:12Z
dc.date.available	2024-02-28T12:06:12Z
dc.date.issued	2023-07-20
dc.identifier.uri	http://ir.bdu.edu.et/handle/123456789/15675
dc.description.abstract	Credit scoring is a method used to assess a customer's reliability and repayment capacity. Currently, Ethiopian financial institutions rely on traditional loan assessment methods, which may lead banks approving risky loans. If financial institutions are unable to manage their loan risks, it may hamper their ability to achieve both their social and financial objectives. To address these challenges, credit scoring can be used as a valuable tool to improve the risk assessment system of credit worthiness’s and improve the performance of financial institutions. This research evaluated the performance of five machine learning models (Decision Tree, Random Forest, Linear Regression, Ridge Regression, and Lasso Regression) to overcome these challenges by using 22 banks dataset with the shape of (321439, 16). The models were evaluated with several metrics including MSE, MAE, and R-squared. The result showed that the RF model had the highest training accuracy 82.63% and the lowest MSE (1312.11) and MAE (29.24) values, indicating better predictive accuracy compared to other models. The model also had the highest R-squared (0.79) value outperforming all other models. The Decision Tree model ranked second in performance. The Linear Regression, Ridge Regression, and Lasso Regression models had similar levels of performance in terms of evaluation metrics.	en_US
dc.language.iso	en_US	en_US
dc.subject	Computer Science	en_US
dc.title	"PREDICTIVE MODELING FOR CREDIT SCORING AND COMPARATIVE DATA ANALYSIS USING MACHINE LEARNING METHODS FOR ETHIOPIAN BANK’S"	en_US
dc.type	Thesis	en_US