FORECASTING DIRECTION AND ANALYZING DENGUE FEVER RISK AREA BY USING GEOGRAPHIC INFORMATION AND DATA MINING IN BANG MAE NANG SUB-DISTRICT, BANG YAI DISTRICT, NONTHABURI PROVINCE

Abstract

Dengue fever remains the major problem with ongoing outbreaks. Each year, outbreak trends are forecasted and forecasted in advance of and during the outbreak. All at the national level, health zone level, provincial level, and district level to prepare and prevent the outbreaks. The researcher is therefore interested in studying the prediction of outbreak trends and analyzing dengue fever risk areas during the outbreak in the area. The purposes of this research are to analyze the factors related to the forecasting direction of the outbreaks and analyze dengue risk areas, predict the direction of the outbreaks and analyze dengue risk areas, and compare the efficiency and accuracy of the forecasting algorithm in Bang Mae Nang Subdistrict, Bang Yai District, Nonthaburi Province. The data in this study are 235 dengue patients then analyze and join spatial data with attribute data, factors selected were 9 factors consisting of (1) Land used (2) Land parcel (3) Household density (4) age (5) Gender (6) Precise (7) Humidity (8) Temperature and (9) Heat index, then to create the models to forecast use 3 algorithms of data mining technique are (1) Decision Tree: CART(Classification and Regression Tree) (2) Decision Tree: ID3 (Iterative Dichotomiser3) and (3) Random Forest. The results show that the top 3 factors for forecasting are the Land used, Land parcel, and Household density. The result of the prediction of the outbreak direction finds that the model can used to predict and analyze risk areas. The model created using the Decision Tree algorithm is the best model and has an accuracy of 98%. The average overall efficiency is 98%.