Reduction of dimensions is a primary goal of multivariate analysis.
Two approaches:
Find an important subset of the original variables.
Synthesize new variables from the original variables.
The creation of new variables can be approached in one of two ways:
projection and mapping.
Projection
Linear combinations of the original variables are created which can
define a new, smaller set of variables, while retaining as much information
as possible. Principal components analysis (PCA) is an example.
Mapping
Mapping transformations are non-linear.
They preserve special properties of the data (e.g. interpoint distances),
while performing data reduction.
The results from mapping can be difficult to interpret.
An Introduction to Chemometrics - Definitions / B A Rock, 469D, 7-1119 /
Jun 5 1997