Limitations in PCA occur mainly due to the previously mentioned main assumptions and the data
at hand. PCA is not a statistical method from the viewpoint that there is no probability
distribution specified for the observations. Therefore it is important to keep in mind that PCA
best serves to represent data in simpler, reduced form.