Analisis Arsitektur Jst Berdasarkan Angka Kematian Akibat Overdosis Narkotika Menggunakan Algoritma Backpropagation

Abstract

Narcotics is a problem in every country, especially in Indonesia. Narcotics are very dangerous because they can damage the nervous system to cause death. Therefore, the development of accurate prediction models is widely developed to determine the mortality rate. This research uses a standard backpropagation artificial neural network to predict drug-induced mortality rates based on relevant historical data. The purpose of this research is to analyze the architecture of the Backpropagation artificial neural network based on the death rate due to drug overdose.  The results of research with 4-10-1 architecture get the best performance 0.00099981 with 100% accuracy. The number of epochs used is 100,000 epochs and finished at epoch 2196 with time 00.07 seconds.