Conduct simulation studies to demonstrate the performance of

supervised principal component analysis in generalized linear models. Some requirements.

$1),$ p $>$ N$,$ i$.$e$.,$ the number of measurements is larger than number

of subjects, $2),$ consider choosing tuning parameters carefully, $3)$ You should

consider alternative methods such as principal component analysis, LASSO or

elastic net, $4)$ consider various simulation settings, $5)$ carefully report your result,

and make comments on your findings.