ESTIMATION OF MISSING VALUES IN REPLICATED FACTORIAL EXPERIMENT
ESTIMATION OF MISSING VALUES IN REPLICATED FACTORIAL EXPERIMENT
No Thumbnail Available
Date
2014
Authors
EKWUEME, CHINENYE LOVELYN
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This study examined the power of Pairwise Deletion (PD), Multiple Imputation (MI) and Expectation Maximization (EM) methods in estimating missing values in cases where the data are missing at random. The data used is a replicated 2 x 3 x 4 factorial experiment in a randomized complete block design (RCBD) and a simulated data set (SMDS) in which data points were randomly selected as missing were used to examine the methods. The result shows that the missing values which are missing at random can be determined using EM method because the estimated values obtained in terms of means, standard error and P-values for all the variables considered were consistent and approximately similar. Also, the results obtained using this method were approximately similar to that of the real-life data set (RLDS) and simulated data set. The study therefore recommends that in a replicated factorial analysis with missing values, EM method has been shown to give better and appropriate results
Description
A THESIS SUBMITTED TO THE POSTGRADUATE SCHOOL, AHMADU BELLO UNIVERSITY, ZARIA NIGERIA IN PARTIAL FULFILMENT FOR THE AWARD OF MASTER OF SCIENCE (M.Sc) DEGREE IN STATISTICS, DEPARTMENT OF MATHEMATICS AHMADU BELLO UNIVERSITY, ZARIA NIGERIA
Keywords
ESTIMATION,, MISSING VALUES,, REPLICATED,, FACTORIAL EXPERIMENT