<![CDATA[Fire occurrence rates in Indonesia increases every year. Fire occurence which increases every year proves that people doesn't really care about the said disaster. Hotspot can be used to identify the event of fire. Fire can be observed through satellite by detecting hotspot on Earth surface. That's why, there is a need for research to predict the number of hotspot which identify fire disaster in a certain time. This research proposes and creates program to predict the number of hotspot occured in Java island using Fuzzy Time Series. The data used is hotspot data in Java island from January 2012 to December 2016. Testing is done to know the accuracy of the number of hotspot prediction in monthly and 10 days period. The best monthly hotspot prediction has MAPE value of 37,128% with the parameter of training data = 80%, testing data = 100%, and the number of interval = 22. The best 10 days period hotspot prediction has MAPE value of 64,4429% with the parameter of training data = 100%, testing data = 20%, and the number of interval = 6. Further research can be done to repair the MAPE from the prediction result.]]>
Copyrights © 2018
