Currently, JMP provides Root Mean PRESS value, Van der Voet T2 and Prob>van der Voet T2 value in Kfold cross validation chart. However, even you repeatedly run the same model (with the number of fold equals to number of observations), the value of Prob>van der Voet T2 will change every times even though the Root Mean PRESS and Van der Voet T2 keep no change. This causes a lot of confusion especially when we want to reduce number of factors for PLS model. Can JMP modify the calculation to keep the Prob>van der Voet T2 change as the same pace of van der Voet T2 change?
See Prob_van der Voet T2 change.jmp.
Zhiwu Liang liang.zw@pg.com Kobe Technical Center Japan
