Dimensionality reduction in data from LASER applications

Abstract

Redundant variables not only in LASER applications, but in all experimental works are disturbing statistical analysis as a result of highly correlation among them. It is not easy sometimes to identify which set of variables is redundant and which one is retained. In addition, consideration of huge sets of variables will make it difficult to point out the joint effects of any subset of variables on a certain phenomenon. It is well know that continuous variables can be transformed into a discrete (categorical) form depending on predefined intervals, thus, the categorical principal component analysis was adopted here in this paper to identify the discarded set of variables when the data contained some variability. The effect of identifying groups of retained variables was compared by observing the natural grouping of elements using single linkage clustering of elements

Keywords

• catpca
• single linkage clustering
• laser