Signal normalization is necessary to bring
signal values into manageable range so as to avoid sharp spikes or values which
may be too close to zero. This helps in stabilizing the algorithm of any
system. Spikes in the dataset can produce unstable results. In PCA, Normalization
is important since it is a variance maximizing exercise. It projects your
original data onto directions which maximize the variance. The first plot below
shows the amount of total variance explained in the different principal
components where we have not normalized the data. As you can see it seems like
only component one explains all the variance in the data.