Determination of dengue hemorrhagic fever disease factors using neural network and genetic algorithms

Abstract

The Aedes aegypti mosquito and the Aedes albopictus mosquito are carriers of the virus that causes Dengue Hemorrhagic Fever (DHF). In Indonesia, the spread of DHF disease has taken place for 41 years. Within this period, there was an increase in the number of spreading areas by 97% and an increase in the number of cases by 99%. Based on the data from previous studies, further information is needed related to the factors that have the most influence on the level of DHF infection in a region. Based on the initial study conducted, there are 6 factors that have the potential to influence the number of DHF cases in an area, namely temperature (X1), rainfall (X2), population density (X3), altitude (X4), distribution of males (X5), and distribution of education level (X6). In this study, the problem of determination dengue disease factors was modeled using a neural network. The activation function in this neural network model then estimated using genetic algorithms. Determination of the best factor is carried out in a genetic algorithm by combining several parameters of the crossover probability (Pc) and mutation probability (Pm). This experiment show that the main factors that influence the spread of DHF in Bandung area are temperature, altitude, distribution of gender, and distribution of education levels. The best accuracy system obtained in this study using these 4 factors reached 72%.