Hi All,
I performed my experiments using RSM.
My R-square, adj R-square and pred-R-square are 0.93, 0.91 & 0.88 respectively which shows model is good fit. But my lack of fit is found to be significant.Validation experiments have been carried out where standard deviation between predicted response and experimental response is found to be 2.9 and 5.95 for 2 different responses.
Can anybody help which is more important lack of fit or R-square & what is the maximum std deviation allowed in validation experiments or optimization runs??