Sumários

Bayesian estimation

27 Maio 2008, 09:30 Jorge dos Santos Salvador Marques

Difficulties of the ML method. The denoising problem.

Use of a priori information. Rescue problem. The a posteriori distribution. MAP and MMSE estimates. Examples.


Classic estimation: linear Gaussian models

20 Maio 2008, 09:30 Jorge dos Santos Salvador Marques

Linear models with additive Gaussian noise. ML estimate of the coefficients. Geometric interpretation. Signal space and noise space. Orthogonal projection of the observations onto the signal space.


Classic estimation: maximum likelihood method

15 Maio 2008, 11:00 Jorge dos Santos Salvador Marques

Asymptotic properties of ML method. Signal processing applications: estimation of the phase of a sinuoid with additive noise.


Classic estimation: maximum likelihood method

13 Maio 2008, 09:30 Jorge dos Santos Salvador Marques

The maximum likelihood (ML) method. Application of the ML method to estimate distributions with unknown parameters from a sequence of iid observations. Identification of signal models (AR model).


Cramer-Rao Bound

8 Maio 2008, 11:00 Jorge dos Santos Salvador Marques

Characterization of estimators. Bias and variance. The Cramer-Rao lower bound. Examples.