SIMPLE INUNDATION EVALUATION SYSTEM BASED ON MACHINE LEARNING USING INUNDATION AND RAINFALL RECORDS IN THE PAST

Abstract

A new simple inundation evaluation method based on the support vector machine, which is one of machine learning methods, using inundation and rainfall records in the past as training data, is proposed. Formulation on the application of the support vector machine to the inundation evaluation is fully described. The system, in which the inundation evaluation method as well as visualizing tools is implemented, has been developed. By comparison of the inundation hazard predicted by this system and the actual inundation that occurred recently, the effectiveness and applicability of this system are discussed.