Penerapan Metode Backpropagation Neural Network untuk Mengidentifikasi Penyakit Cacar Monyet

Abstract

Monkeypox is a zoonotic disease caused by the monkeypox virus of the genus orthopoxvirus, which belongs to the family of poxviridae and is considered one of the dangerous skin diseases. Previously, the disease was detected using a PCR testing of skin lesion samples and analysis of the patients clinical symptoms. However, the increasing global spread of monkeypox in non-endemic regions demands for a rapid and accurate diagnostic method. This research proposes a machine learning approach based on artificial neural networks, employing the Backpropagation Neural Network (BPNN) method for monkeypox classification. The research scenario was conducted with variations in dataset split ratios (70:30, 80:20, and 90:10), one hidden layer, 18 neurons in the hidden layer, a learning rate of 0.1 and 0.01, and the application of ReLU and Binary Sigmoid activation functions, and compare of test results between the data balancing method SMOTE with the original dataset. The best scenario results were obtained from testing on the original dataset with a data split configuration of 80:20, 500 epochs, learning rate 0.1, achieving an accuracy of approximately 70.36%, a precision of 72.33%, a recall of 88.08%, and an F1-score of 79.26%.