Discriminant non-stationary signal features’ clustering using hard and fuzzy cluster labeling

Abstract

<p>Current approaches to improve the pattern recognition performance mainly focus on either extracting non-stationary</p> <p>and discriminant features of each class, or employing complex and nonlinear feature classifiers. However, little</p> <p>attention has been paid to the integration of these two approaches. Combining non-stationary feature analysis with</p> <p>complex feature classifiers, this article presents a novel direction to enhance the discriminatory power of pattern</p> <p>recognition methods. This approach, which is based on a fusion of non-stationary feature analysis with clustering</p> <p>techniques, proposes an algorithm to adaptively identify the feature vectors according to their importance in</p> <p>representing the patterns of discrimination. Non-stationary feature vectors are extracted using a non-stationary</p> <p>method based on time–frequency distribution and non-negative matrix factorization. The clustering algorithms</p> <p>including the K-means and self-organizing tree maps are utilized as unsupervised clustering methods followed by a</p> <p>supervised labeling. Two labeling methods are introduced: hard and fuzzy labeling. The article covers in detail the</p> <p>formulation of the proposed discriminant feature clustering method. Experiments performed with pathological</p> <p>speech classification, T-wave alternans evaluation from the surface electrocardiogram, audio scene analysis, and</p> <p>telemonitoring of Parkinson’s disease problems produced desirable results. The outcome demonstrates the benefits</p> <p>of non-stationary feature fusion with clustering methods for complex data analysis where existing approaches do not</p> <p>exhibit a high performance.</p>