Association analysis of earthquake distribution in Indonesia for spatial risk mapping

Abstract

Indonesia is an earthquake-prone country which surrounded by tectonic plate boundary and ring of fire area. In 2016, there are 14 times of earthquake rates ≥ 5 Richter on average per month occurred in Indonesia. Because the high rates of earthquake in Indonesia connected in earthquake tectonic plate boundary, it is important to analyze a causal-effect relationship between earthquake hit in several regions. This paper proposes a new system for causal-effect relationship analysis of earthquake data distribution in Indonesia. This system presented an automatic spatio-temporal cluster-based earthquake data distribution and developed an association-mining of the data projected to provinces for risk-mapping of region in Indonesia. The system has 3 main features: (1) Subspace Earthquake Data Selection, (2) Automatic Spatio-Temporal Clustering, (3) Association Mining for Earthquake Data Distribution, (4) Causal-Effect Relationship Visualization, and (5) Risk-Mapping Earthquake Analysis. We applied our system with seismic data in Indonesia taken from 1963–2016. The results of our experiment was found the interesting patterns relation from the association of earthquake distribution in Indonesia. Provinces with strong relation are Maluku, North Maluku and North Sulawesi that always appear as a rule in every experiments period and give each other the risk of the earthquake.