Modeling and classification of biological signals

Abstract

This thesis examines a number of marine biological signals and the problem of modeling by autoregressive techniques using a prony-svd algorithm to accurately represent segments of biological signals. Two methods are employed to classify the biological signals from the model parameters. The first classification method is based on a Neural Network implementation using a commercial software package. The second method is accomplished by using a distance measure, based on spectral ratios, with respect to modeled reference signals.