LINEAR REGRESSION

Shivali Srivastava
Institute for Signal and Information Processing
Mississippi State University, Mississippi State, MS 39762
email: shivali@isip.msstate.edu

ABSTRACT

The technique of regression, in particular linear regression, is the most popular statistical tool. There are several forms of regression such as linear, non-linear, simple, multiple, parametric and non-parametric.The major purpose of regression is to explore the dependence of one variable on other. One is called as response variable and other is known to be predictor variable. The goal is to forecast the values of the dependent variables on the basis of one or, more explanatory variables by setting up an appropriate probability model. The type of model depends on the types of variables involved.

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: