Dengue outbreak prediction based on meteorological data using machine learning techniques in Kota Bharu / Nor Farisha Muhamad Krishnan

Dengue fever is a mosquito-borne infection that causes a high temperature, rashes, severe headache, muscle and joint discomfort, pain behind the eyes, and, in rare cases, bleeding. Rainfall, humidity, temperature, precipitation, floods, human movement, population, and the environment are only a few of the elements that induce dengue fever, including climatic and non-meteorological elements. This study used two different datasets that are dengue data and meteorological data that aims to identify the significant meteorological variables then develop a machine learning model to predict the dengue outbreak and proposed the machine learning. Dengue outbreak can be defined as 2 or more number of reported dengue cases in 7 days in certain regions. Random Forest feature selection is used to identify the significant meteorological attributes. It showed that maximum temperature, minimum temperature, average humidity and rainfall are significant for predicting dengue outbreaks. For modelling, Artificial Neural Network (ANN) and Decision Tree (DT) model were used to predict the dengue outbreak. Both models undergo parameter tuning to optimize the model. For ANN the different number of hidden nodes and decay were used to improve the model while for DT, maximum depth and complexity parameter were varying to improve the model. Both models, ANN and DT are evaluated based on accuracy, sensitivity and specificity showing that ANN (Accuracy = 69.05%, Sensitivity = 98.84%, Specificity = 3.80%), performed better than DT (Accuracy = 67.46%, Sensitivity = 97.11%, Specificity = 2.53%). The government and Vector Borne Disease Control (VBDC) may have preventive measures to handle the dengue outbreak as the meteorological parameters affect the dengue outbreak.