Abstract of Thesis Presented to the Graduate School
 of the University of Florida in Partial Fulfillment of the
 Requirements for the Degree of Master of Science

 NON-LINEAR NEURAL NETWORKS FOR
 MODELING AND ONLINE SEGMENTATION OF
 NON-STATIONARY VENTILATORY SIGNALS

 By

 Helene L. Chini

 December 2002

Chair: Jose C. Principe
Major Department: Electrical and Computer Engineering

 Many real-world time series are multi-modal, where the underlying data generating

process switches among several different subprocesses. The difficulty in analyzing such

systems is to discover the underlying switching process, which entails identifying the

number of subprocesses, the dynamics of each subprocess and the switching pattern

among them. Unfortunately, real world time series usually are non-stationary. The goal of

this work is to perform online segmentation of a real-world time series, i.e., tracking of

the slow evolution as well as detecting the sudden changes. Several algorithms in the

framework of gated experts are implemented and tested on biomedical data, using non-

linear neural models.

 These approaches are first tested on a synthetic data set, to understand the effect of

the system and model parameters, then on real data, yielding encouraging results.