Eigenvalue Decomposition of Matrix that doesn't fit in memory

I have a Matrix, X, that is too large to fit into memory. I would like to perform an eigenvector & eigenvalue decomposition of X. My question: is it possible to take a large matrix X, decompose it into many smaller matrices: {x1, x2, ....xn} perform some equivalent of the eigen-decomposition of these smaller matrices and then finally recombine their resultants into what is the equivalent eigen-decomposition of X?

Thanks!

Answer 1

[Leah]

Not really, maybe principle component analysis is a better option.