Description

About the Book

The book Statistical Criteria for Selection of Functional Forms in Regression Analysis presents a comprehensive and rigorous treatment of regression modeling with special emphasis on choosing appropriate functional forms and models. It begins with the fundamental theory of linear regression, explaining classical assumptions, estimation through ordinary and restricted least squares, and the consequences of assumption violations. The authors carefully develop the mathematical foundations needed for regression analysis, including matrix methods and the Kronecker product, to handle complex systems of equations. A major focus of the book is on criteria for selecting regressors and models, where methods based on mean square error, prediction accuracy, and information criteria such as AIC, BIC, and related measures are discussed. The text also emphasizes diagnostic testing, residual analysis, and model adequacy to ensure reliable inference. Beyond linear models, the book explores nonlinear regression, alternative functional forms such as logarithmic, polynomial, exponential, and Cobb–Douglas models, and advanced systems like Seemingly Unrelated Regression Equations (SURE). New methodologies for selecting the best functional form using goodness-of-fit measures and feasible GLS estimation are proposed and evaluated. Overall, the book bridges theory and application, offering a valuable reference for postgraduate students, researchers, and practitioners interested in accurate regression model specification and selection.

About the Author 

Dr. K. Naga Vihari, M.Sc., B.Ed., M.Phil., Ph.D.
Associate Professor
Department of MBA
International Institute of Business Studies
Bengaluru North-562110.
Mail ID: viharinaga@gmail.com
viharinaga@iibsonline.com

“Dr. M. Naresh, M.Sc., Ph.D.
Assistant Professor(C)
Department of Statistics
Sri Venkateswara University
Tirupati-517502.
Mail ID: naresh.ygr@gmail.com