Multivariate data occur in almost all branches of science. Whenever
more than one attribute is measured for each experimental unit, the
data is termed as "multivariate". The statistical methods used to
analyze such data are called multivariate statistical
techniques. These are very useful for researchers who strive to make
sense of large, complicated and complex data sets.One fundamental
distinction between multivariate methods is that some are classified
as "variable directed techniques" while others are classified as
"individual directed techniques". The former are primarily concerned
with relationships that might exist among the response variables being
measured while the latter are primarily concerned with relationships
that might exist among the experimental units and/or individuals being
measured.
During the course of this talk, we will explore and analyze some
common multivariate techniques used in the field of speech
recognition. Principal Components Analysis (PCA) is an example of a
"variable directed technique" that is very useful for analyzing speech
data. Linear Discriminant Analysis (LDA) is an example of a
"individual directed technique" that is also very useful. We will
construct small real life examples and apply these techniques and
compare the results. We will also talk about the applications of these
methods for classification purposes. Independent Components Analysis
(ICA) is a relatively new multivariate technique that is gaining
importance in the speech research community. We shall also try to
analyze this approach and compare it with the traditional techniques
like LDA and PCA.
Additional items of interest: