A method for dam health diagnosis based on the improved attribute reduction

Abstract

Abstract. Since the traditional rough set theory can easily result in information loss during attribute discretization and its attribute reduction is too complex, the fuzzy rough set theory and the golden section method are introduced for dam health diagnosis. With attribute fuzzification replacing attribute discretization, and attribute significance as a condition of attribute reduction, the dam health rough set diagnosis model is improved. Next, the improved dam health rough set diagnosis model is applied to a practical project. Results show that the improved attribute reduction put forward in this paper can more fully demonstrate factors influencing uncertainty of the dam health status. The diagnosis results, while more reasonably reflecting the dam's practical health status, can provide a new research path for dam health diagnosis.