Abstract:It is one of the key issues that making knowledge discovery effectively in a Cognitive Radio(CR) engine design. Basing on the research about Rough Set Theory and C4.5 algorithm of decision tree, this study presented a model of CR knowledge discovery designed by combination of rough set and decision methods and studied its feasibility through a case. Using data based on simulation platform of MATLAB 802.11a physical layer as CR perception sample, decision tree sequence was trained, and decision tree was built for knowledge extraction. Then the accuracy and performance of the design model was evaluated by confusion matrix. The simulation results show that the proposed design model gets high classification accuracy rate, can enhance the interpretability of knowledge，and therefore has preliminarily achieved the purpose of knowledge discovery for cognitive radio and learning from the experiences.