Analisis dan Prediksi Default Kartu Kredit dengan Model Machine Learning

Abstract

This research aims to study dataset partitioning, prediction, and model selection with the best performance in the case of credit card default classification. The trained models include logistic regression, k-nearest neighbor, artificial neural networks, decision trees, support vector machines, random forests, and gradient boosting. The objective is to determine the best model and the most suitable model for credit card default classification. The research involves dataset visualization, dataset preprocessing, model selection, parameter tuning for each model, and using machine learning models to predict the test dataset. Dataset visualization is conducted to understand dataset characteristics and identify patterns among features. Dataset preprocessing is performed to handle missing values, feature normalization or standardization, outlier removal, and encoding or transformation of categorical features. Model selection is carried out to choose the most appropriate machine learning model for credit card default classification. Parameter tuning for each model aims to improve model performance and generalization. Finally, the trained machine learning models are utilized to predict the test dataset and provide the final evaluation of model performance, specifically the F1 score.