A Geographic Information System for Hotspot Occurrences Classification in Riau Province Indonesia

Abstract Forest fire is a serious problem because it causes damages to forest ecosystem and surrounding environment in many areas in Sumatera and Kalimantan including Riau Province. One of activities in forest fire prevention is to provide relevant information about hotspot and characteristics of areas where hotspots probably occur. Such information can be presented in a geographic information system (GIS) for classifying hotspot occurrences. This study aims to develop a classification module for hotspot occurrences in a GIS using OpenGeo Suite 3.0. The classification module was made using the decision tree method namely the C4.5 algorithm. Applying the C4.5 algorithm on the forest fire data generated 109 classification rules with accuracy of 69.56%. The SIG has major features such as displaying map, pan map, zoom, in, zoom out, scale, mouse position and a classification module for hotspot occurrences based on characteristics of areas. The GIS enables users to determine whether an area is potential for hotspot incidence or not based on its characteristics. Potential hotspot occurrences can be used in decision making related to forest fire prevention.