Cluster Analysis Algorithms for RS and WWLLN Data Processing

Abstract

Thunderstorm activity is indirectly taken into account when using statistics on forest fires. Another way is to use the data of the lightning discharge finding sensors. This chapter proposes using the data of the World Wide Lightning Location Network (WWLLN). To assess the probability of forest fire occurrence, it is necessary to know the energy and the spatial distribution of lightning discharges. For the analysis of the data of the storm-direction-finding WWLLN, it is proposed to use clustering algorithms. For the computational experiment, the region covering the Timiryazevskiy forestry of the Tomsk region (55.93 - 56.86) on the north, (83.94 - 85.07), and the data on lightning discharges registered by the WWLLN network in this region in July 2014. The sample data were 273 lightning discharges. The results of clustering are presented, as well as the image of the silhouette index for each object, the average value of the ASW index for grouping solutions obtained using the complete, kmeans, and dbscan algorithms.