Abstract :
We report unsupervised classification of past ten years seismic events occurred in India based on their spatial
location and magnitude using k-means clustering. We useearthquake dataset derived from European-Mediterranean
Seismological Center in this work.With the proliferation of efficient analysis, clustering is a time-honored and well
understood process for extracting meaningful patterns from large incoherent data sets. We demonstrate the k-means
clustering as a tool for analyzing seismic datasetstogether with visualization for interpreting the results. This work
exhibits performance of k-means clustering by tuning number of clusters and iterations. Performance is evaluated
with reference to within clusters sum of square errors. The present investigation derives six earthquake clusters and
thus depicts that k-means has the potential to exhibit the unsurpassed tool for earthquake cluster analysis.
Keyword :
k-means, clustering, earthquake, European-Mediterranean Seismological Center.