Parametric sensitivity in building fuzzy decision trees: an experimental analysis

Fuzzy decision tree (FDT) induction is an extraction technique of fuzzy rules, which has been widely used in handling ambiguous classification problems related to human's thought and sense. The entire process of building FDT is based on a specified parameter (called significant level) which seriously affects the computation of fuzzy entropy and classification result of FDT. Since the value of this parameter is usually given in terms of human experience while building a FDT, it is very difficult to determine its optimal value. This paper attempts to give some guidelines of how to automatically choose the optimal value of this parameter by analyzing the analytic expression between this parameter and fuzzy entropy and further by investigating the decision trees sensitivity to the parameter perturbation.