Fig. 1 is illustrating mean squared error for both the
methods as a function of data dimensionality. In the figure
‘+’ sign indicates PCA using EVD method and ‘.’ sign indicates
Fast PCA method. It can be observed that the errors
generated by both the methods are very close to each other
for all the data dimensions. From this it can be inferred
that the difference in performance of these methods in
terms of the mean squared error is negligible.