Outlier Detection on Hotspot Data in Riau Province using OPTICS Algorithm
N L Febriana, Imas Sukaesih Sitanggang · IOP Conference Series Earth and Environmental Science · 2017
Hotspot is a point that shows the coordinate of an area with relatively higher temperature than another surrounding area. Hotspot is one of indicators of forest and land fires. An effort in forest and land fires prevention is detecting hotspot occurrences and its outlier. The purpose of this study is to detect outlier occurrences on hotspot data in Riau Province using the density based algorithm namely Ordering Point to Identify the Clustering Structure (OPTICS). The data used are hotspots in Riau Province for the period of 2001 to 2012. In order to find the best clustering results, OPTICS was executed on the parameter Eps of 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.1 and MinPts of 1 to 6. This study found that outliers are commonly occurred in hotspot data in 2007 at the parameter Eps of 0.01 and MinPts of 6. This study identifies 906 outliers in hotspot data in 2007 with SSE of clustering result of 0.0219. Outliers are mostly found in February spreading on several districts including Siak district in Riau Province.