On Improving the Performance of Logistic Regression Analysis Via Extreme Ranking

Document Type

Contribution to Book

Publication Date

8-11-2020

Publication Title

Computational and Methodological Statistics and Biostatistics

DOI

10.1007/978-3-030-42196-0_15

Abstract

Logistic regression models for dichotomous or ordinal dependent variables is one of the generalized linear models. They have been frequently applied in several fields. In this chapter, we present more efficient and powerful performance of the logistic regression models analysis when a modified extreme ranked set sampling (modified ERSS) or moving extreme ranked set sampling (MERSS) are used and further improving the performance when a modified Double extreme ranked set sampling (modified DERSS) is used. We propose that ranking could be performed based on an available and easy to rank auxiliary variable which is associated with the response variable. Analytically and through simulations, we showed the superiority performance of the logistics regression analysis when modified ERSS, MERSS, and DERSS are used compared with using the simple random sample (SRS). For illustration purposes of the procedures developed, we use a real dataset from 2011/12 National Survey of Children’s Health (NSCH).

Comments

Copyright belongs to Springer. Information regarding the dissemination and usage of journal articles can be accessed through the following link.

Share

COinS