RGU organises lecture on ‘multivariate analysis’

RONO HILLS, 23 May: The Centre for Development Studies (CDS) of Rajiv Gandhi University’s (RGU) economics department organised a special lecture on ‘multivariate analysis’ on Wednesday as part of its lecture series programme.

West Bengal-based University of Burdwan’s Economics Professor Pravat Kumar Kuri delivered a lecture on ‘Model selection in multivariate analysis,with special reference to principal component analysis’, the university informed in a release.

In his lecture, Prof Kuri said, “Economic studies are concerned with the analysis of relationship among the variables, and there are instances when a researcher has to deal with more than two variables.

“Multivariate analysis provides techniques and methods for dealing with simultaneous analysis of associations among more than two variables,” he said.

He presented the various models for multivariate analysis, and described the situations under which such models are applied for the analysis of data, the release said, adding that “Prof Kuri also showed different charts, flow diagram and equations to explain the procedure of multivariate analysis.”

He informed the scholars that one of the most popular and widely used methods of multivariate analysis is the principal component analysis (PCA). “When the variables are highly correlated and interdependent,there would be a problem of multi-collinearity, and so regression analysis cannot be applied. Under such a situation, the PCA can be applied to examine the relationship by computing the principal component,”he added.

Prof Kuri informed that “PCA is nowadays used in computation of human development index as this method helps to find out the weights of each dimensions.

“Earlier, human development index was computed by assigning equal weight to each dimension. But this procedure has come under severe criticism as various dimensions are interrelated, and so there was a need to obtain data-driven weights, which is provided by the PCA,” he added.

“One can obtain as many principal components as the number of variables, but the number of principal components to be retained would be decided on the basis of Eigen values,” he said.

Lastly, he showed the steps to be followed in the SPPS software for PCA, the release stated.

RGU Economics HoD Prof SK Nayak and CDS coordinator Prof Vandana Upadhyay also spoke, it said, adding that the lecture was attended by postgraduate students, research scholars, and faculty members of the economics department.