Decision rule extraction for total amount type of debris flow based on rough set theory

This paper proposes a method using rough set theory to extract decision rules for the observation data of 30 debris flow emergencies in Jiangjiagou Ravine, Yunnan, China. First, the row-column discretization is adopted in data discretization process, and which also reduces some condition attributes. Then, The attributes reduction work is done based on Niche Genetic Algorithm. Finally, a attribute value reduction operation based on discriminated matrix is utilized. The result shows the relationship between the total amount type of debris flow and the rainfall characteristics including indirect antecedent rainfall, direct antecedent rainfall, one hour antecedent rainfall and process rainfall. The acquired decision rules can provide help for decision maker to process some emergencies, which may decrease the hazard of debris flow and the treatment scope of emergency.