ACADEMICIA: An International Multidisciplinary Research Journal

  • Year: 2014
  • Volume: 4
  • Issue: 5

A study on the variable importance in regression models – partial standardization technique

  • Author:
  • M. Sudarsana Rao, M. Venkataramanaiah, M. Subbarayudu
  • Total Page Count: 12
  • DOI:
  • Page Number: 167 to 178

*Research Scholar, Department of Statistics, S.V. University, Tirupati, India

**Professor, Department of Statistics, S.V.University, Tirupati, India

Abstract

Variable importance means the importance of different independent variables considered in the model understudy. It aims at investigating the contribution of each independent variable. But in some cases, it means not only to judge the contribution of particular variable, but also, it helps in the selection of a best subject of independent variables. In this article an attempt is made to trace out the relative importance of explanatory variables in multiple regression by employing standadardization technique and partial standardization technique. Lastly, comparative study has been made between the standardization technique and partial standardization technique to see the responsiveness of the respective techniques in tracing out relative importance of explanatory variables in regression models.