Data Mining and Artificial Intelligence: 2017

Suyitno, The Houw Liong , Arif Budiman

ABSTRACT

Malaria still remains a public health problem in developing countries and changing environmental and weather factors pose the biggest challenge in fighting against the scourge of malaria. Malaria is an endemic disease in most of Indonesian area, especially in rural and remote areas. The incidence and spreading of malaria were influenced by environmental and weather factors, namely temperature, rainfall, humidity and length of daylight. Therefore this study would like to developed a malaria incidence prediction system based on environmental and weather factors, so that it may assist Indonesian Ministry of Health to control malaria. The method used to solve this problem was Hopfield Neural Network.

Hopfield Neural Network method have being application for malaria forecast because this method can give the recurrent malaria classification. This weather substance in Hopfield method as the neuron input and then the result of simulation process will be recurrent as input until reach stabil condition. The best performance while predicting malaria incidence in the year of July 2008 – December 2009, was accuracy 94.14%, and MAPE 5.86%. Using the training dataset is 80% from total data, with a mean threshold data.

Using a Hopfield Network can reduce the number of iterations to get the convergence toward the target pattern. In our study to get an output that converges on average takes 8 iterations within several seconds.

Keywords: Malaria, Prediction, Artificial Neural Network, Discrete Hopfield.