May 14, 2013
I am doing PLS in JMP using leave one out cross validation and vander Voet T2 statistic as the criteria to decide on the number of factors to keep in the model.
The roblem I have is with interpreting the cross validation RMSE values. So, I understand that the variables are transformed to standard normal distribution before the PLS is performed.
Then, I was thinking the cross validation RMSE values in the output can be converted to the scale of the variables by simply multiplying it by the standard deviation (SD) of the observed
response variable since the standard normal distribution has mean 0 and SD of 1. Am I thinking correctly or is there a different way to convert this RMSE to the scale of the variables?
Second question is about getting the calibration RMSE values? How can I do this?
Any help would be greatly appreciated. There is no information about this in JMP documents.