A method for augmenting experimental data to maximize [X′X] is developed. The rule is: scan a list of candidates and find the conditions in the experimental space where the variance of a predicted response is the greatest and obtain the next run at those conditions. The method is applicable for any form of the model. Illustrations are included.
No takes yet. Share an insight, caveat, or question.
Otto Dykstra (1971) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: