Pakistan Science Abstracts
Choose Theme
1
2
3
4
Theme 1
— Green & Gold
Print
Save as PDF
Export BibTeX
Close
No Detail Found!
Home
›
Pakistan Journal of Statistics
Design and analysis of left censored data from highly fractionated experiments.
Author(s)
Abdul H. Chowdhury
Department of Mathematics and Statistics, University of Windsor Windsor, ON, N9A 4Z5 Canada
Nasser S. Fard
Department of Mechanical, Industrial and Manufacturing Engineering Northeastern University, Boston, MA, 02115 U.S.A
Abstract
A method is developed for analyzing censored data, where the experiments involve several factors and the response data is subject to censoring. Many practical experiments involve a large number of factors in a small number of runs for highly fractionated experiments (FIFE). Standard methodologies for analyzing censored data are computationally complicated and often are not feasible. A regression-based technique was adopted for imputation of censored data. Initial model selection and model fitting are determined from an orthogonal setup of the whole experiment. This method deals with censoring in the response variable and could be used in other applications where multiple regression or least squares procedure is applicable. The proposed procedure is demonstrated through reanalyzing censored data from a real industrial experiment.
Publication Details
Page(s)
159-171
DOI
DOI not available
Published
Journal: Pakistan Journal of Statistics, Volume: 20, Issue: 1, Year: 2004
Request the full text from PASTIC
More from
Pakistan Journal of Statistics
Interval estimation of the availability of a two-unit standby system with non-instantaneous switch-over and `dead time`.
Estimation of threshold parameter using ranked set sampling procedure.
Regularization parameter selection and inference for statistical inverse problems.
Simple accurate unique: the methods of modern likelihood theory.
A simple general procedure for unequal probability sampling without replacement.
0
Citations
0
Downloads
41
Views