Pruning pada Fuzzy Decision Tree dalam Klasifikasi Data Iklim dan Titik Api di Daerah Tjilik Riwut, Palangkaraya, Kalimantan Selatan
Abstract
Forest fire is influenced by several factors, such as humidity, solar radiation intensity, regional temperature, and rainfall. This research aimed at finding the information and knowledge from hotspot and climate data, especially those four attributes. The research data was taken from Tjilik Riwut, Palangkaraya, South Kalimantan in year 2001-2004. Data mining technique used for extracting the information and knowledge is classification using decision tree method. In this research, fuzzy aproach is adapted to solve uncertainty of data. To improve the accuracy of classification process, pruning tree method is utilized. Tree that has the highest accuracy is converted to be the rule. The formed rule shows that the amount of hotspot is inversely proportional with the scale of humidity. This research also proves that pruning process in a tree can improve the accuracy of classification process.
Collections
- UT - Computer Science [2236]