- FA vs PCA
- FA tries to explain the covariance and/or correlation
structure whereas PCA tries to account for the
variability structure among the measured variables
- Both of them perform transformations on the correlation
matrix
- FA assumes an underlying model
- PCA is a special case of the normal FA procedure
(principal factoring technique)
- FA vs MDS
- FA and MDS are both useful for clustering purposes
- When we have the actual data points, it is preferrable
to perform FA
- When we have the distances between the data points
(rather than the data points), performing a MDS is
preferred
- We shall illustrate the above points by way of a small
example