EE 5352. STATISTICAL SIGNAL PROCESSING. 3 Hours.
Estimation of correlation functions, power spectra, and probability density functions using training data. Algorithm development using maximum likelihood and Bayes approaches. Lower bounds on estimates of error variance. Optimal feature extraction. Design of approximate Bayes discriminants for classification using training data. Feature selection using subsetting and transformation approaches. Branch and bound. Applications in bioengineering, remote sensing, and telecom.