Analisis Data Deret Waktu Hotspot Provinsi Riau Berdasarkan Tipe Penutupan Lahan

Abstract

Riau is a province in Indonesia which has high occurences of forest fires and also causes polemics to its neighboring countries. Fire occurences identification can be performed by processing hotspot distribution data. In this research, decomposition of the time series data was used to identify hotspot occurence patterns. Decomposition of time series data shows several patterns of hotspot over the year 2001 to 2012. In addition, classification methods namely conditional inference tree and k-nearest neighbor were applied to create classifiers for hotspot data based on land cover types. The highest accuracy of classifier is 33.3% for the dataset with seven target classes of land cover types in monthly period.