<![CDATA[Forest / land wildfire is one of the disasters that occur every year in some countries in the world. This incident got more attention from the government because it caused many losses both in the economic, ecological, and social. Indonesia is a country with a high rate of forest / land wildfire disasters. Indonesia suffered losses of up to Rp 209 trillion by 2015. As a result of losses incurred an early prevention is needed, which one can be done by grouping areas with potential forest fires by utilizing hotspot data. Forest wildfires are marked by the detection of fire spots by satellites indicated as hot spots. This research uses hotspot data with parameter of latitude, longitude, brightness, frp (fire radiative power), and confidence by using K-Medoids method. K-Medoids method is a clustering method that serves to split the dataset into groups. The advantages of this method is able to resolve the weakness of K-Means method that is sensitive to outlier. The result of this research shows that the use of K-Medoids method can be used for the process of hot spot data clustering with the best silhouette coefficient in amount of 0.56745 on the use of 2 clusters by using 7352 data. The results of the clustering analysis showed that using 2 clusters resulted in a group of data with the potential of high potential with an average brightness of 344.470K with average confidence of 87.18% and medium potential with average brightness of 318.800K with Average confidence of 58.73%.]]>
Copyrights © 2017
