PERAMALAN JUMLAH KORBAN DEMAM BERDARAH DENGUE MENGGUNAKAN METODE SEASONAL AUTOREGRESSIVE INTEGRATED MOVING AVERAGE (SARIMA) DAN NEURAL NETWORK

• Main Authors: , Maria Bellaniar, , Adhistya Erna Permanasari, S.T., M.T., Ph.D.
• Format: Theses and Dissertations NonPeerReviewed
• Published: [Yogyakarta] : Universitas Gadjah Mada 2014
• Subjects: ETD
• Online Access: https://repository.ugm.ac.id/133370/; http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=74002
• Institution: Universitas Gadjah Mada

Dengue fever is transmitted by Aedes aegypti and it is considered as dangerous disease due to its number of victims, which is first rank in ASEAN and second rank in the world. The number of dengue victims in Charitas Hospital Palembang tends to increase in certain months and indeterminate in every month. In addition, the data of the dengue victims are not used as an evaluation to reduce the number of victims. It becomes the basic of forecasting for the number of dengue victims in the next year, so the Charitas Hospital Palembang is able to reduce the number of dengue victims in the future. The research to predict the number of dengue patients has been done by using various techniques of artificial intelligence and statistical method. This research is associated with forecasting number of dengue fever patients using time series of Charitas Hospital Palembang over the last 10 years. SARIMA (Seasonal Autoregressive Integrated Moving Average) and Neural Network were used in the forecasting. SARIMA represents statistical model whereas Neural Network represents artificial model. The obtained result of significant pattern of Dengue Fever begins in December, reaches the summit in January, begins to decline in February and March followed by forecast number for each month in the next year. The best model of SARIMA is (0,1,0) (0,1,1) and Neural Network with 12 input layers, 28 hidden neurons, and 1 output layer. The value errors of SARIMA are MSE (1602.04), RMSE (40), MAPE (27.46%), and MAD (27.48) whereas for the Neural Network are MSE (240.21), RMSE (15), MAPE (12.59%), and MAD (10.64). Based on the obtained error value, the more appropriate method to get more accurate forecasting results is Neural Network because the obtained error value is lower than SARIMA.