03-06-2014 12:32 PM
I am using the PLS procedure of SAS to stablish relation among some geographical and climatic variables with frequecies os 11 mutations in a gene related with animal adaptation to environmental conditions.
I need to know the standard deviations of the bj coefficient (s(bj) for scaled and centered data. To obtain this I need to gain access to the vector of n bij coefficients obtained by leave-one-out cross validation jackknifing (i=1....n) n= number of observations. Reading the SAS procedures manual I can not identify how to obtain this output and if this is possible.
Someone knows how to do this?
Thank you in advance
03-06-2014 01:56 PM
I don't think PROC PLS provides this information.
You'd have to write your own version of PLS in PROC IML and then do the leave-one-out cross validation there and save the resulting coefficients.