Colloquium

The Department of Statistics
presents

Ta-Hsin Li
Department of Statistics
Texas A&M University

Filter Bank Methods for Time Series Analysis

Abstract

Filter banks have been used to suppress unwanted noise, enhance desired signals, and decompose signals into different components. When coupled with suitable statistical parameters, filter banks can also be used to characterize statistical properties of time series. This talk discusses several methods of time series characterization using filter banks. Examples are given to demonstrate these methods in a number of applications including speech analysis, nondestructive testing, radar imaging, and time series forecasting.