Using Data Mining Techniques for the Management of Seismic Vulnerability
Natural hazards, and especially earthquakes, are often recurring phenomena. Therefore, there is a permanent need for solutions to reduce earthquake losses by developing technologies, procedures, knowledge, and tools for seismic design and rehabilitation of buildings and infrastructure. A key point to an effective decision making process that aims at mitigating their effects is building a model of the underlying facts. A Geographical Information System (GIS) is a framework able to assemble, keep, process and display specific information, identified by geographical location, which can combine layers of information to give the user a better understanding about that location. By using a Geographical Information System containing geospatial data, one can develop useful scenarios to reduce natural disaster risk and vulnerability of structures. In this paper, we describe a way of applying data mining techniques from the artificial intelligence field to earthquake analysis in order to make a better investigation of the available data. These methods are capable of finding “hidden” correlations among different subsets of data, which cannot be revealed by means of simple statistics.