Feasibility of Ambiguity Analysis for Binary Classification from Non-Stationary Signals

Non-stationary signal analysis for classification of signals is the recent area of research. Various researchers have contributed significantly in this field. However the inherent randomness in the non-stationary signals is still a challenge to be scaled. In this work, Framelet and ambiguity analysis based technique for classification of non-stationary signal into two categories is proposed. The proposed technique proves to be an efficient method for binary classification from signals which have bursts and signals with saw teeth. Sensitivity of the proposed technique is 96% and the specificity is 94 %.