COMPARATIVE ANALYSIS OF MACHINE LEARNING ALGORITHMS WITH PRINCIPAL COMPONENT ANALYSIS FOR BREAST CANCER PREDICTION

Abstract

Breast cancer is the leading cause of cancer death in Indonesia and globally, making early detection crucial. This study aims to compare the performance of Machine Learning algorithms (K-Nearest Neighbor/KNN, Support Vector Machine/SVM, Logistic Regression, Random Forest, and Decision Tree) in predicting breast cancer types using the Wisconsin Breast Cancer dataset. This dataset consists of 569 samples and is divided into training and testing sets using an 80:20 ratio. Evaluation methods include accuracy, precision, recall, F1-Score, and dimensionality reduction impact analysis using Principal Component Analysis (PCA). This study contributes by providing a comparative evaluation of several machine learning algorithms before and after PCA implementation. The results show that Random Forest achieved the highest accuracy (0.97) before PCA implementation but experienced a moderate decrease (0.94) after dimensionality reduction. KNN showed the highest consistency, maintaining a stable accuracy of 0.94 both before and after PCA. This study concludes that Random Forest is the most effective algorithm when using the full feature set, while KNN shows greater stability after dimensionality reduction, thus this study recommends KNN as a reliable solution to support accurate and efficient breast cancer diagnosis and early detection.