摘要
Learning Vector Quantization(LVQ)originally proposed by Kohonen(1989)is aneurally-inspired classifier which pays attention to approximating the optimal Bayes decisionboundaries associated with a classification task.With respect to several defects of LVQ2 algorithmstudied in this paper,some‘soft’competition schemes such as‘majority voting’scheme andcredibility calculation are proposed for improving the ability of classification as well as the learningspeed.Meanwhile,the probabilities of winning are introduced into the corrections for referencevectors in the‘soft’competition.In contrast with the conventional sequential learning technique,a novel parallel learning technique is developed to perform LVQ2 procedure.Experimental resultsof speech recognition show that these new approaches can lead to better performance as comparedwith the conventional
Learning Vector Quantization(LVQ)originally proposed by Kohonen(1989)is a neurally-inspired classifier which pays attention to approximating the optimal Bayes decision boundaries associated with a classification task.With respect to several defects of LVQ2 algorithm studied in this paper,some‘soft’competition schemes such as‘majority voting’scheme and credibility calculation are proposed for improving the ability of classification as well as the learning speed.Meanwhile,the probabilities of winning are introduced into the corrections for reference vectors in the‘soft’competition.In contrast with the conventional sequential learning technique, a novel parallel learning technique is developed to perform LVQ2 procedure.Experimental results of speech recognition show that these new approaches can lead to better performance as compared with the conventional LVQ2