An improved online learning algorithm for general fuzzy min-max neural network

This paper proposes an improved version of the current online learning\nalgorithm for a general fuzzy min-max neural network (GFMM) to tackle existing\nissues concerning expansion and contraction steps as well as the way of dealing\nwith unseen data located on decision boundaries. These drawbacks lower its\nclassification performance, so an improved algorithm is proposed in this study\nto address the above limitations. The proposed approach does not use the\ncontraction process for overlapping hyperboxes, which is more likely to\nincrease the error rate as shown in the literature. The empirical results\nindicated the improvement in the classification accuracy and stability of the\nproposed method compared to the original version and other fuzzy min-max\nclassifiers. In order to reduce the sensitivity to the training samples\npresentation order of this new on-line learning algorithm, a simple ensemble\nmethod is also proposed.\n