Classification Model Disease Risk Areas Endemicity Dengue Fever Outbreak based Prediction of Patients, Death, IR and CFR using Forecasting Techniques

Abstract

model time series data becomes a process of model use to create prediction (forecast) to an event in the future based on past event which is known as data. The data gives benefit to the interrelatedness among factors that can be the predictor variables. Deep investigation to time series data will give information to find out the data pattern. One of the examples of data series is dengue fever phenomena. Dengue fever is a contagious disease which can become an endemic. The spread of dengue fever needs a particular supervision from related party under the supervision of health department. Dengue fever control is used to avoid the disease to stop spreading. It is done to give important information, so the society can know endemic area from a particular social environment. The research is aimed to do improvisation toward dengue fever surveillance system by conducting a forecast of dengue fever occurrences in a particular area. Forecast technique was conducted with three (3) algorithm approaches. The test of three algorithm forecasting was done to predict dengue fever occurrences in 12 months and to help the activity to control spreading cases of dengue fever to stop spreading in the area of Yogyakarta. Forecast experiment approach from three methods in which linear regression, multilayer perceptron, and sequential minimal optimization regression (SMOreg) showed that linear regression had better accuracy compared to sequential minimal optimization regression (SMOreg) and multilayer perceptron.