In this paper, we propose FRID (Fuzzy-Rough Interactive Dichotomizes); a methodology for the induction of decision trees using rough set based measures to capture cognitive uncertainties inherent to databases. These measures are: 1) Fuzzy-roughness and 2) Fuzzy-rough entropy. Developed FRID algorithms have been initially applied to various real-world benchmark datasets, and experimentally compared with the three fuzzy decision tree generation algorithms reported so far. Simulation results confirm that the use of proposed strategy leads to smaller decision trees and as a result better generalization performance.