Estimation of representative truncated orthogonal subplans of complete factor experiment plan
AbstractA criterion for estimation of the quality of a truncated orthogonal subplan of the complete factor experiment plan is introduced, which characterizes the level of mixing effects of the factors and their interactions. A method for calculation the degree of representativeness for truncated orthogonal plans is proposed. Taking into account of the criterion makes it possible to improve the estimate accuracy for multifactor regression under the conditions of small selection.
Mathematical methods, models, problems and technologies for complex systems research