ANALISIS PERFORMA ALGORITMA XGBOOST PADA KLASIFIKASI KANKER PARU-PARU

Abstract

This study aims to analyse the performance of the Extreme Gradient Boosting (XGBoost) algorithm in lung cancer classification. This test used a dataset obtained from Kaggle, comprising 220,632 data points and 23 attributes. The research methods included exploration, data preprocessing such as cleaning, selection, and transformation, as well as dividing the data for training and testing five times (70%:30%, 75%:25%, 80%:20%, 85%:15%, 90%:10%). Model evaluation metrics such as accuracy, precision, recall, and F1-score were used in conjunction with the confusion matrix. Based on the results of XGBoost model testing with varying parameter settings, it demonstrates good performance even when the proportion of training and testing data is altered, achieving the highest accuracy of 95.76%, precision of 96%, recall of 100%, and F1-score of 98%.