Supervised wavelet-network based fuzzy-logic classifier performance on the UCI databases

Supervised machine learning is an important field with many immediate applications. As a result, there is an increasing number of public tools with a diversity of learning approaches. In this paper we propose a new architecture of wavelet network classifier learnt by a fast wavelet transform (FWN). This classifier is well suited for data classification and has many advantages compared to other ones. We have contributed by proposing a new classification way. It is characterized by its novel technique for processing data similarity distances, with involvement of a fuzzy decision support system (FDSS) in decision-making, which operates a human reasoning mode. The empirical results demonstrate that the proposed system outperforms the other ones, published in the literature, in terms of global classification rates on different well known datasets.