Given that principal components analysis (PCA) have been performed on a dataset, the question is what does the distances between the coordinates of the data points in the transformed (PCA) space mean? For instance, if Euclidean distance is calculated between point A and point B on the PCA coordinates is this perhaps equivalent to Mahalanobis distance between point A and point B with the original coordinates?
NOTE: assuming that the data is scaled as part of the PCA analysis. The code in R to perform the PCA step would look like this (where df
is the dataframe containing the original records)
pca <- prcomp(df, scale=TRUE)