AN EARLY-WARNING MODEL OF DAM SAFETY BASED ON ROUGH SET THEORY AND SUPPORT VECTOR MACHINE

摘要

There are strong non-linear and dynamic relations between dam behavior and its influence factors. The early-warning models of dam safety need to be described with non-linear function. Rough set theory is used to implement data pretreatment on dam safety monitoring. Main factors influencing dam safety are mined. An early-warning model is built with support vector machine. The proposed model can provide an effective performance of approximation and forecast for the relations between dam behavior and above mined factors. The system rule on dam behaviors can be learned and induced from the prototype observations of dam safety. The expression and parameters of early-warning model need not to be predefined.