Penerapan Struktur Backpropagation Pada Jaringan Syaraf Tiruan Untuk Mendeteksi Gangguan Penyakit Tropis

Abstract

 Tropical disease is the most common diseases in tropical and subtropical regions. Many factors affected the spread of these diseases, such as poor sanitation and bad environment. Islam establishes the principles in maintaining health through the cleanliness, wudu, and taking bath regularly. Technology through the expert system development tried to transform the expertise knowledge into computers that can mimic the workings of the human brain. One of the methods applied is Artificial Neural Network (ANN) with backpropagation structure. This method detected the tropical diseases of patients, including Dengue Hemorrhagic Fever (DHF) and Typhoid Fever to perform the appropriate treatment as early as possible. ANN diagnosed the type of diseases by identifying the pattern of symptoms in patients. ANN training was presented using 80% of training data and 20% test data. The binary sigmoid activation function [0 1] is used. The learning rate (α) values 0.05, 0.1, 0.2, 0.5, 0.75 and the hidden layers values 10, 50 and 100 are used in testing process. ANN trained the input symptoms, thus the results proposed whether patients affected by any kinds of tropical disease or not.

Keywords: DBD, hidden layer, JST, learning rate, Tifoid