Some novel classifiers designed using prototypes extracted by a new schemebased on self-organizing feature map

We propose two new comprehensive schemes for designing prototype-based clas sifiers. The scheme addresses all major issues (number of prototypes, gener ation of prototypes, and utilization of the prototypes) involved in the des ign of a prototype-based classifier. First we use Kohonen's self-organizing feature map (SOFM) algorithm to produce a minimum number (equal to the num ber of classes) of initial prototypes. Then we use a dynamic prototype gene ration and tuning algorithm (DYNAGEN) involving merging, splitting, deletin g, and retraining of the prototypes to generate an adequate number of usefu l prototypes. These prototypes are used to design a "1 nearest multiple pro totype (1-NMP)" classifier. Though the classifier performs quite well, it c annot reasonably deal with large variation of variance among the data from different classes. To overcome this deficiency we design a "1 most similar prototype (1-MSP)" classifier. We use the prototypes generated by the SOFM- based DYNAGEN algorithm and associate with each of them a zone of influence . A norm (Euclidean)-induced similarity measure is used for this. The proto types and their zones of influence are fine-tuned by minimizing an error fu nction. Both classifiers are trained and tested using several data sets, an d a consistent improvement in performance of the latter over the former has been observed. We also compared our classifiers with some benchmark result s available in the literature.