Applied Multivariate Analysis

The course will comprise of 33 modules spread over 12 week with the 7th week assigned for the mid-term assessments. The final assessment will be held after the 12th week.

In the course we discuss various multivariate techniques which are basic to the study of multi-dimensional data. These techniques can be applied to diverse fields of natural, biological and social sciences. The course is thus intended for students and researchers who studies involve the simultaneous analysis of several variables 

Multivariate methods are primarily of two types. The first is the extension of existing univariate techniques to the multivariate set-up. This brings in more complexity, but the advantage is that the accounting of inter-relationships between the variables leads to better inferences. The second type of studies is unique to multivariate data only. These are either too trivial or the problems are non-existent for univariate data.

In this discourse we will deal with both these types. The Multivariate Linear Models, including multi-response regression, MANOVA and MANCOVA belong to the first category, while the Dimensional Reduction techniques, which include principal component methods and factor analysis techniques, and Clustering and Classification techniques, which include cluster analysis and discriminant analysis belong to the latter.   

Of course to fully understand applied multivariate techniques, it is often necessary to have a reasonable knowledge of the theory of multivariate analysis. Part of the course is devoted to an introductory study of basic multivariate theory, particularly those related to the multivariate normal distribution.