New criteria for robust integer-valued designs in linear models

• Published in: Computational statistics & data analysis Vol. 51; no. 2; pp. 723 - 736
• Main Authors: Adewale, Adeniyi J., Wiens, Douglas P.
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 15.11.2006; Elsevier Science; Elsevier
• Series: Computational Statistics & Data Analysis
• Subjects: Bias-constrained; Contamination; D-optimal design; Exact sciences and technology; Experimental design; Finite design space; General topics; Mathematics; Minave design; Misspecification; Multivariate analysis; Numerical analysis; Numerical analysis. Scientific computation; Numerical methods in probability and statistics; Polynomial regression; Probability and statistics; Sciences and techniques of general use; Simulated annealing; Statistics; Finite design space; Minave design; Primary 62K05; Bias-constrained; Misspecification; Simulated annealing; D-optimal design; 62F35; secondary 62J05; Contamination; Polynomial regression; Error estimation; Bias; Estimator robustness; Robust estimation; Cluster model; Linear model; Variance; Primary 62K05; 62F35; secondary 62J05; Misspecification model; Bad specification; Data analysis; Approximation; Linear regression; Prediction; Mean estimation; Heteroscedasticity; Integer; Statistical regression; Extrapolation; Statistical computation; Regression model; Experimental design; Bias-constrained; Contamination; D-optimal design; Finite design space; Minave design; Misspecification; Polynomial regression; Simulated annealing; Mean error; Biased estimation

We investigate the problem of designing for linear regression models, when the assumed model form is only an approximation to an unknown true model, using two novel approaches. The first is based on a notion of averaging of the mean-squared error of predictions over a neighbourhood of contaminating...