IS PRINCIPAL COMPONENT ANALYSIS MORE EFFICIENT TO DETECT DIFFERENCES ON BIOMECHANICAL VARIABLES BETWEEN GROUPS?

Abstract

The biomechanical analysis investigates variables such as angles, inter-segmental forces and moments at the joints. When the relevant parameters (e.g., range of motion, peak values) are selected a priori from these variables, they could not perfectly represent the information content of the original dataset. Therefore, in this study we want to validate the efficacy of the Principal Component Analysis (PCA) in overcoming the limitations of the a priori selection of the parameters. An application study is reported; the lower-limb joint mechanics between patients operated with two different surgical techniques for a total hip arthroplasty are analyzed with both the traditional analysis and the PCA. The findings from the two methods converged, but the PCA identified new sources of variability not previously detected.