Genetic Algorithm Based K-Means Fast Learning Artificial Neural Network

摘要

The K-means Fast Learning Artificial Neural Network (KFLANN) is a small neural network bearing two types of parameters, the tolerance, δ and the vigilance, μ. In previous papers, it was shown that the KFLANN was capable of fast and accurate assimilation of data. However, it was still an unsolved issue to determine the suitable values for δ and μ in [12]. This paper continues to follows-up by introducing Genetic Algorithms as a possible solution for searching through the parameter space to effectively and efficiently extract suitable values to δ and μ. It is also able to determine significant factors that help achieve accurate clustering. Experimental results are presented to illustrate the hybrid GA-KFLANN ability using available test data.