Neuro-computing Method for Data Mining

Among the advantages of the cerebellar model articulation controller neural network (CMAC NN) include very fast learning, reasonable generalization capability and robust noise resistance, explaining why CMAC NNs are conventionally used in robot control. This study considers the feasibility of CMAC NN as an efficient data mining (DM) method, indicating that the CMAC NN can extend its network topology flexibly to achieve DM applications. Therefore, this study introduces CMAC NN for applying classification and time series prediction problems. The solved problem, network topology, learning algorithm and recommended parameter settings are described as well. Results of this study contribute to efforts to extend network topology for the CMAC NN in DM applications.