| dc.description.abstract | Principal component analysis (PCA) is a dimension-reducing tool that replaces the variables in a multivariate dataset by a smaller number of derived variables. Dimension reduction is often undertaken to help in interpreting the data set but, as each principal component usually involves all the original variables, interpretation of a PCA can still be difficult. One way to overcome this difficulty is to select a subset of the original variables and use this subset to approximate the data. Procrustes analysis as a measure of similarity, is used to measure the efficiencies of the alternative variable selection methods in extracting representative variables Because of its unavailability in statistical software, a package program, using Mathematica 8.0, is composed for variable selection. There are four variable selection methods, based on PCA and procrustes analysis, which have been described and examined along with different criteria levels for deciding on the number of variables to retain in the analysis. The methods are B2, B4, procrustes analysis in the principal component score, and procrustes analysis method. The result show that variable selection programs B2 as the best variable selection method followed by procrustes analysis method. Moreover, it is found that all of the methods considered give the measure of efficiency of more than 99.04%. | en |
