Graphical Method for Evaluating the Prediction Variance Characteristics of Composite Mixed-Resolution Design

Authors

Polycarp E. Chigbu, Department of Statistics, University of Nigeria, Nsukka, Nigeria
Cynthia N. Umegwuagu
Eugene C. Ukaegbu

Author Country (or Countries)

Nigeria

Abstract

In this work, the prediction variance characteristics of the Composite Mixed-Resolution Designs (CMRD) were explored in spherical region using graphical procedures. Three-dimensional Variance Dispersion Graphs (VDG) was proposed and used as a graphical technique that assesses the designs’ prediction capabilities. The replication of the cube and star portions of the CMRD as well as the centre point was considered in the evaluation of the designs. Under the spherical design region, the VDGs display the designs’ prediction variances throughout the design region. The results show that replicating the star portion of the composite mixed-resolution designs for the practical axial distance, in most cases improves the scaled prediction variance property of the design.

Digital Object Identifier (DOI)

http://dx.doi.org/10.18576/jsap/090204

Recommended Citation

E. Chigbu, Polycarp; N. Umegwuagu, Cynthia; and C. Ukaegbu, Eugene (2020) "Graphical Method for Evaluating the Prediction Variance Characteristics of Composite Mixed-Resolution Design," Journal of Statistics Applications & Probability: Vol. 9: Iss. 2, Article 4.
DOI: http://dx.doi.org/10.18576/jsap/090204
Available at: https://digitalcommons.aaru.edu.jo/jsap/vol9/iss2/4