Influences of Perturbations of Training Pattern Pairs on Fuzzy Morphological Neural Networks
Yang Jing-yu · Control Engineering of China · 2011
Abstract When constructing the neural networks,training pattern pairs obtained by a certain acquisition way always have perturbations.How to measure the perturbations,a new measure method of perturbations of fuzzy sets is first proposed in the paper.Two kinds of morphological associative memory networks had the same very good fault-tolerant abilities of erosive/dilative noise were analyzed by many scholors.But the investigation reveals that two kinds of networks have widely different to perturbations of training pattern pairs based on the measure method.There is the good one to the perturbation of training pattern pairs in the fuzzy morphological associative memory networks,the other is relatively bad.The work in the paper is of some benefit to the performance analyses,the choice of learning algorithms,and the guidance to training pattern pairs acquisition for fuzzy morphological associative memory networks