Dear Vignesh

The Lack-of-fit is designed to test if the model proposed fits adequately the recorded datas (or if the model represents accurately the studied phenomena) so one can use it for prediction. So if the lack-of-fit is not shown, the model is not useful.

Regards!

NOBOSSE

Is this the case even R-seq is 0.99 and most of P-value for all variables and their interactions were

Lack of fit should be not significant.........dont select the model manually..incase of those issues LOF may not be shown......anyhow without considering u cant able to proceed your work