Similarity Classifier With Weighted Ordered Weighted Averaging Operator

Abstract

In this paper we present a similarity-based classifier that utilizes a weighted ordered weighted averaging (WOWA) operator in the aggregation of infor-mation. The aggregation process used in the WOWA operator is studied and tested with five different Regular Increasing Monotonic (RIM) weight generators or quantifiers. The proposed approach is tested with five real-world data sets. For comparison purposes the obtained results are compared to results from two previously introduced classifiers. The proposed new classifier showed comparatively improved performance over for all studied data sets. The results indicate that there are benefits in using a WOWA operator in similarity classifiers.