- Factor Analysis is a variable directed technique
- Unlike PCA, FA depends on a reasonable statistical underlying
model to explain the data distribution
- While PCA is concerned with explaining the variability in the
data, FA is concerned with explaining the covariance and/or
correlation structure among the measured variables
- Loosely formulated, FA determines whether p response
variables can be partitioned into, say, m subsets,
each consisting of a group of variables tending to be more
highly related to others within the subset than to those in
other subsets
- A more rigorous mathematical formulation follows