The talk will contain the different models to represent linear
regression such as mathematical model, statistical model and the
probabilistic model. The emphasis will be given to least square
criterion of regression. I will also discuss some fundamental aspects
of "maximum likelihood linear regression (MLLR)" as a transformation
approach to speaker adaptation. MLLR is a method that transforms
mixture components that are transformed similarly and introduces the
concept of regression classes as a set of mixture components that are
transformed similarly.
Additional items of interest: