Talk:Statistics/Multivariate Data Analysis/Principal Component Analysis

Please whatever computational aspect we do here let us focus on MatLab...as it seems the language most people use these days to understand the mathematical concepts as much easier to learn and comprehend and less numerical using graphics.

Variance
The main point of PCA is that the components explain, in order, the variance of the data. If we start off with n components and do a PCA, we finish up with n components. However, the first principal component explains the most variance in the data, the second component the next most variance, and so on. This means that it's often possible to omit the least important components -- dimensionality reduction.

All the principal components are linear combinations of the original components, so nothing gets added or lost (up to the point where we discard unimportant principal components).

--84.9.67.105 (talk) 15:18, 11 April 2008 (UTC)