Abstract
In this study, the authors have experimentally established an assessment rule of debris flow torrents applicable to degree of disaster risk. For the first trial, the representing hazard data were selected by eliminating surrounding noises as much as possible by means of SVM method. Second, the disaster risks were classified respectively by using separating hyperplane constructed by the selected representing hazard data. Then, each hazard rank was studied and analyzed by application of rough set theory, so that the assessment rule has been established. The rule through this process has high versatility, and has a good consistency with actual events of disaster occurrence or non-occurrence. From the method above mentioned, the results of this study, show that the most appropriate rule can be selected from survey data of any debris flow torrents and the system can be applicable to utilize for the judgment of hazard rank, as well.
