Using the steepest slope menu
This help file describes usage of the steepest slope menu.
This help file describes usage of the steepest slope menu for response surface models. The menu is only available if there is an rsm model available with at least one response.
It is necessary to select the model (as rsm models cannot be active models in R Commander). The default degree is 2, but 1 is also possible.
The checkbox provides a switch between classical steepest ascent in first order models (start in the middle) or models that are almost linear with far out optima (start in a saddle point). The latter option is reasonable for second order models, for which the Hessian reveals that there is still room for improvement = the stationary point is a nearby saddle point, it is more appropriate to try increasing (decreasing) the response starting at that saddle point and walking into both directions.
Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.