Debris flow risk assessment using principal components analysis and rough set techniques

Prediction of natural disasters and their consequences is difficult due to the uncertainties and complexity of multiple related factors. Thus, developing a method of debris flows assessment seems necessary. Previous researchers have proposed several methods, such as regression analysis, fuzzy mathematics, and artificial neural networks for debris-flow hazard assessment. However, these methods need further improvements to eliminate the high relativity in their results. For improving the assessment of debris flows risk, a hybrid assessment model combined principal components analysis and rough set theory to determine hazard levels of debris flow in regions, with steps like determining hazard-level-type regions. Firstly, we use principal components analysis technique to reduce debris flows data, and attain a data set of attribute reduction. Secondly, we utilize rough set technique to classify the reduction set, and obtain a classification about debris flows risk assessment. Our proposed methodology was then illustrated to assess the regional debris hazard of Yunnan Province in China. The results indicate that the proposed debris flows risk assessment model is accurate and efficient, and can improve the comparability and reliability of the assessment to some degree.