A novel data mining technique of analysis and classification for landslide problems

Landslides during earthquakes have led to severe casualties and have resulted in damaged structures and facilities. The goal of the present study is to analyze the landslide problems in a remote area-Shei-Pa National Park in Taiwan. Spatial information techniques (Remote Sensing and Geographic Information System) with an innovative data mining technique, Discrete Rough Set (DRS) method, are incorporated to our study for analyzing landslides, their distribution, and classification. The present study provides how to find (1) the most representative data of landslide samples from the existing database, (2) the core attributes of the target categories: Normalized Difference Vegetation Index (NDVI) and Vegetation Index (VI), and (3) the thresholds (segment points) of each attribute on the target categories. A conventional approach, C4.5 Decision Tree Analysis, is used as a comparison. The methodology discussed in this study is of help to the analysis of landslide problems and thus facilitates the informed decision-making process.