| http://www.w3.org/ns/prov#value | - where Y is an nxq matrix of proxies, U is an nxp matrix of temperature indicators or PCs, G is a pxq matrix of coefficients determined in your first equation, E is the nxq matrix of errors. (I???ve suppressed the constant term, but that should be in there as well.) Your second equation estimates a reconstructed value (or values) of U, say U*, from new values of the proxies, say Y*, under the assum
|
