Non-stationary signal analysis using temporal clustering

• Published in: Neural Networks for Signal Processing VIII. Proceedings of the 1998 IEEE Signal Processing Society Workshop (Cat. No.98TH8378) pp. 304 - 312
• Main Authors: Policker, S., Geva, A.B.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 1998
• Subjects: Clustering algorithms; Data mining; Hidden Markov models; Probability distribution; Random number generation; Random processes; Signal analysis; Signal processing algorithms; Speech analysis; Switches
• ISBN: 078035060X; 9780780350601
• ISSN: 1089-3555; 2379-2329
• DOI: 10.1109/NNSP.1998.710660

We present a model of nonstationary time series generated by switching between a finite number of random processes and apply temporal clustering to estimate the model's parameters. Applications of the algorithm to segmentation of nonstationary time series and a simple example of preprocessing a speech signal will be discussed. The model defines a nonstationary composite source generated by randomly switching between elements of a finite number of random processes. The switching probability distribution which underlies the behavior of the switch is controlled by a time varying vector of parameters which is used to determine a different switching probability in each time instant. This definition allows us to analyze a drift between disjoint states of the composite model.