Data Clustering and Zonationof Earthquake Building Damage Hazard Area Using FKCN and Kriging Algorithm

Abstract

The objective of this research is to construct the zonation of earthquake building damage hazard area using fuzzy kohonen clustering network (FKCN) algorithm for data clustering and kriging algorithm for data interpolation. Data used consists of the earth data in the form of peak ground acceleration (PGA), lithology and topographic zones and Iris plant database for algorithm validation. This research is comprised into three steps which are data normalization, data clustering and data interpolation using FKCN and kriging algorithm and the construction of zonation. Clusterization produces three classes of building damage hazard data. The first class is consisting of medium PGA,dominantby high compaction lithology in the topography of inland area. The second class with low PGA, dominant low compaction lithology in the lowland topographic zone and the third class with high PGA, dominant by un-very low compactionlithology in swamp topographic zone. Banda Aceh cityas location sample is divided into three building damage hazard zone which is high hazard zone, medium hazard zone and low hazard zone for building damage which is located towards inland area.