Graphical Methods and Principles for Data Analysis Ⅱ: Graphically Comparing Many Curves Using PCA

摘要

Principal components analysis (PCA) is a standard multivariate statistical analysis technique that can be adapted to curve characterization problems when there are a large number of curves. The mathematics behind PCA is somewhat difficult, requiring a good background in linear algebra. There is, however, a straightforward geometric interpretation that is at the heart of how PCA works, so we can leave the complicated mathemat-ics to the PCA software and focus on applying it without losing any essential information.