Sunday, December 4, 2011

How is an eigenvector decomposition conducted with a mixture of continuous, categorical and binary variables?

Trying to conduct an eigenvector decompsition on a correlation matrix that contains categorical, continuous and binary variables. I am trying to reduce the number of varibles I use for a regression, and am trying to identify the principle components.|||Try to decompose one eigenvector linearly and then extend it by cofactoring the matrix.

No comments:

Post a Comment