<![CDATA[Determining appropriate credit limits is essential for financial institutions to manage credit risk effectively while optimizing revenue. This study aims to develop a predictive model for credit limits using linear regression, incorporating primary features such as Rating, Income, and Balance. The dataset consists of 400 credit card customer records with 11 variables, comprising both numerical and categorical data. The research follows the Cross-Industry Standard Process for Data Mining (CRISP-DM) methodology, covering stages including business understanding, data understanding, data preparation, modeling, evaluation, and deployment. Data analysis was conducted using Google Colab, involving quality assessment, categorical feature encoding through label encoding, and data normalization utilizing MinMaxScaler. Correlation analysis results indicated that Rating, Income, and Balance have strong correlations with Credit Limit, hence these three variables were chosen as primary predictors for the modeling process. ]]>
Copyrights © 2025
