Hotspot Occurrences Classification using Decision Tree Method

Abstract

Appllcation of geospatJal and data mining techniques in forest fires research have resulted Interesting and useful Information in decision making related to the forest fires management. This paper presents a result of the study in appl)lng the C4.5 algorilhm on a forest fire datllset in the Rokao Hilir distrk1, Rlau Pro~lnce, Indonesia. The data!ct consists of hotspot occurrence locations. human activity factors, and land cover types. Human activity factors include city center locations, roads network and rh'crs network. The results were a decision tree which contaJns 18 lea\es and 26 nodes \\ith accuracy about 63. 17~• . Most of positive examples (the area "ith hotspot occurrences) and negatin examples (no hotspot occurrences In the area) that arc Incorrectly classified by the model are located near rivers and roads.