Statistical Inference and Confidence Intervals for Absolute Partial Mean Dominance

By : P. Bansal , S. Arora , K.K. Mahajan , Sangeeta Arora

Page No: 223-238

Abstract
Absolute Partial Mean (APM) is related to Bonferroni curve in the same way as Absolute Lorenz curve with Lorenz curve. The curve gives the idea of absolute inequality and can be shown to have relation with all the main curves of inequality viz. Lorenz curve, Generalized Lorenz curve, Bonferroni curve, Absolute Lorenz curve and Comic curve. The present paper considers the asymptotic variance – covariance structure of APM curve ordinates at some pre-selected points. Asymptotically non – parametric joint confidence intervals for APM curve ordinates are derived. Testing for APM dominance is carried out through two different approaches viz. Confidence Interval Approach and Specified Interval Approach and a comparison of these two testing procedures is done by comparing their corresponding powers. A real life illustration is also included.

Authors :
P. Bansal, S. Arora and K.K. Mahajan : Department of Statistics, Panjab University, Chandigarh – 160014, India
Sangeeta Arora : Professor, Department of Statistics, Panjab University, Chandigarh-160014, India