ANALISIS PERBANDINGAN ALGORITMA SUPPORT VECTOR MACHINE DAN ARTIFICIAL NEURAL NETWORK DALAM PREDIKSI DAN KLASIFIKASI KUALITAS PISANG

Abstract

This research to compare the performance of the Support Vector Machine (SVM) for banana quality classification and Artificial Neural Network (ANN) algorithms in predicting banana quality. The Research method involves using the RBF kernel for SVM and the MLP architecture for ANN. The dataset on banana quality consists of 8000 samples and 8 columns, used to for this research dataset. SVM and ANN are trained and evaluated using the same datasets to compare their accuracy. The average experimental results show that SVM (RBF Kernel) achieves higher accuracy, with attributes such as accuracy, precision, recall, and f1-score at 0,981 compared to ANN (MLP) which only achieves 0,98. However, ANN shows a tendency to be more flexible in handling data variations. SVM with the RBF kernel is more suitable for modeling banana quality with a stable dataset, while ANN with MLP is more appropriate for datasets with more complex variations.