APPLICATION OF PCA FOR THE CLASSIFICATION OF HUMAN WHOLE BLOOD SAMPLES

Principal component analysis (PCA) is applied as a powerful tool to identify the possible correlations of different elements and patterns in large collection of blood samples. PCA has been successfully applied to analytical data of Cu, Cd, Li, Mg, Pb and Zn for their possible correlations with each other in 500 blood samples of healthy human subjects. The graphical representation of scores has been used to conceive the relative disparity in large collection of blood samples, while the loadings have been used to explain their any possible relationship among elements. Loadings show a direct correlation between Pb and Cd and negative correlation between Cu and Zn in the blood samples.The scores suggest 12 samples as outliers. These samples were rejected from the normal blood samples. However, the scores of the rejected blood samples indicate no defined correlation between Pb and Cd and inverse correlation between Cu –Zn, Mg -Cd and Cd - Li.