Fuzzy Design Of Nearest Prototype Classifier

In pattern classification problems, many works have been carried out with the aim of designing good classifiers from different perspectives. These works achieve very good results in many domains. However, in general they are very dependent on some crucial parameters involved in the design. An alternative is to use fuzzy relations to eliminate thresholds and make the development of classifiers more flexible. In this paper, a new method for solving data classification problems based on prototypes is proposed. Using fuzzy similarity relations for the granulation of the universe, similarity classes are generated and a prototype is built for each similarity class. In the new approach we replace the relation of similarity between two objects by a binary fuzzy relation, which quantifies the strength of the relationship in a range of [0; 1]. Experimental results show that the performance of our method is superior to other methods.