Stepwise multiple linear regression approaches built multistep models step by step. The following procedures are followed to build a model through stepwise multiple regressions:
(i) an initial model is recognized
(ii) by altering the descriptor or variable combination, a previous step is repeated with the stepping crite-rion of F = 4 for inclusion and F = 3.99 for exclusion.
(iii) when a specified maximum number of steps has been reached or/and when stepping is no longer possible with the given stepping criterion, the search is terminated.