PENERAPAN ALGORITME BACKPROPAGATION NEURAL NETWORK UNTUK ESTIMASI JUMLAH KASUS DBD BERDASARKAN DATA CUACA

Abstract

Dengue fever is widespread throughout the tropics which tends to have a seasonal pattern, namely before and after the rainy season. Infection is caused by one of the closely related dengue viruses, commonly called a serotype, which causes mild symptoms to symptoms that require medical treatment and hospitalization, even death can occur if the case is severe. Based on surveillance data, the number of cases in 2022 will be 3,190 people. One of the efforts to reduce the incidence of DHF is by forecasting the incidence of DHF to prevent an increase in DHF cases which continues every year. This research was forecasted using the independent variables average temperature, average humidity, average rainfall, and wind speed. The data used is public through surveillance and the BMKG website and the data used is data from 2018 to 2022. In this study using the backpropagation neural network algorithm, the model used is 4-3-1, where there are 4 variables in the input layer, 3 units in the hidden layer, 1 unit in the output layer with a learning rate value of 0.04, and momentum of 0.09 and the results are RMSE 4,347.