Analisis Perbandingan Algoritma Klasifikasi Decision Tree, K-Nearest Neighbors, Naive Bayes, dan Random Forest pada Data Pemilihan Legislatif KPU Menggunakan Kurva ROC

Abstract

In the context of the digital information era, analysis of general election data is crucial for understanding political dynamics. Legislative election data from the Indonesian General Election Commission (KPU) provides insight into voter behavior and election results. Selection of an appropriate classification algorithm is the main challenge in producing accurate predictions. This study compares four classification algorithms: Decision Tree, K-Nearest Neighbors (KNN), Naive Bayes, and Random Forest, using Receiver Operating Characteristic (ROC) curves as the main evaluation. The results show Random Forest performs best in handling legislative election data, providing important insights for future policy and research.