Machine Learning-Based Credit Risk Classification Using German Credit Data

Credit risk assessment is an important process in financial institutions to minimize potential losses due to non-performing loans. This study aims to apply a machine learning approach in classifying credit risk using the German Credit Dataset. The research method includes exploratory data analysis, data preprocessing, and the application of the Logistic Regression algorithm as a classification method. The dataset consists of numerical and categorical attributes that represent the financial and demographic characteristics of the credit applicant. The test results showed that the built model produced an accuracy rate of 66% on the test data, with a recall value for the high-risk class of 68%. These results suggest that machine learning approaches can be used as a decision support system in credit risk assessment, although further development is still needed to improve classification performance.