Integration of decision tree and fuzzy rule induction on landslide susceptibility modeling

This study tries to integrate decision tree and fuzzy rule induction algorithms (DTFRI) for landslide susceptibility modeling based on rainfall-induced and shallow landslide events. Eleven geospatial factors are considered, including topographic, vegetative, environmental, geological and man-made information. The landslide inventory and factors are overlapped to obtain the training data for modeling and verification. In general, two strategies are utilized for the model verification, i.e. space- and time-robustness. The former is to separate samples into training and check data based on a single event. The latter is to predict (classify) later landslide events with a landslide susceptibility model which is constructed from earlier events. In this study, the constructed landslide susceptibility model derived from the DTFRI algorithm is applied to classify samples and verified by the time-robustness method and the results is also compared with the decision tree classifier. Experimental results indicate that the decision tree classifier can reach high classification accuracy based on the space-robustness strategy but has poor performance to predict (classify) consequent events. However, the DTFRI algorithm can significantly improve the prediction (classification) accuracy in the test cases.