Perbandingan Performa Algoritma CNN, SVM, dan KNN pada Klasifikasi Citra Deteksi Kurangnya Perhatian Pengemudi

Driving is an activity that is commonly done by people today. One of the causes of accidents while driving is driver negligence. Driver behavior can be predicted using machine learning and deep learning models. This research has compared 3 models namely Convolutional Neural Network (CNN), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN) in the classification of driver behavior images. This research uses a dataset containing 14859 images consisting of 985 test data, 11952 training data, and 1922 validation data and divided into 6 classes. This research produced a CNN model with 96% training accuracy, 90% validation accuracy, 90% precision, 90% recall, and 90% f1-score. Then the SVM model produces 85% training accuracy, 84% validation accuracy, 84.4% precision, 84.4% recall, and 84.4% f1-score. Last, the KNN model produces 91% training accuracy, 86% validation accuracy, 86.5% precision, 86.5% recall, and 86.5% f1-score.