Research on influencing factors reduction method of ship maintenance cost based on fuzzy rough set

Influencing factors reduction plays an important role in ship maintenance cost forecasting. Recently, the grey correlativity analysis method is often used to deal with the problem. But the reducts by this method is not satisfied. In this paper, we attempt to use fuzzy rough set theory to solve the problem. Under the framework of fuzzy rough set theory, the definitions of fuzzy equivalence relation, discernibility power and relative reduct are presented. Based on these definitions, influencing factors reduction model is established. Considering the case that the decision attribute is continuous, we redefine general discernibility matrix. Based on improved general discernibility matrix, a new heuristic algorithm for reduction is constructed. The practical experiment shows that the reduct found by the proposed method gets a better performance compared with the grey correlativity analysis method.