Abstract
In designing an experiment, we often assume a model and then find a best design satisfying one or more optimal properties under the assumed model. This approach works well when we are absolutely sure about the assumed model that it will fit the experimental data adequately. In reality, we are rarely sure about a particular model in terms of its effectiveness in describing the data adequately. However, we are normally sure about a set of possible models that would describe the data adequately and one of them would possibly describe the data better than the other models in the set. The pioneering work of Srivastava [33] introduced the search linear model with the purpose of searching for and identifying the correct model from a set of possible models that includes the correct model. This paper focuses on model identification through the use of the search linear models, particularly in addressing the fundamental issues and important challenges in statistical design and analysis of experiments while presenting an overview on this area of research. Two important research areas developed over time using the search linear models are in factorial designs and row-column designs.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
H. Akaike, Fitting autoregressive models for prediction, Ann. Inst. Staist. Math., 21 (1969), 243–247.
H. Akaike, Information theory and an extension of the maximum likelihood principle, in: 2nd International Symposium on Information Theory (ed: B. N. Petrov and F. Csaki), Akadémiai Kiadó (Budapest, 1973), 267–281.
D. A. Anderson and A. M. Thomas, Weakly resolvable IV.3 search designs for the p n factorial experiments, J. Statist. plann. Inference, 4 (1980), 299–312.
K. Chatterjee, Search designs for searching for one among the two-and three-factor interaction effects in the general symmetric and asymmetric factroials, Ann. Inst. Statist. Math., 42 (1990), 783–803.
K. Chatterjee and R. Mukerjee, Some search designs for symmetric and asymmetric factorials, J. Statist. plann., Inference, 13 (1986), 357–363.
K. Chatterjee and R. Mukerjee, Search designs for estimating main effects and searching several two-factor interactions in general factorials, J. Statist. Plann. Inference, 37 (1993), 385–392.
S. Ghosh, On main effect plus one plans for 2m factorials, Ann. Statist., 8 (1980), 922–930.
S. Ghosh, On some new search designs for 2m factorial experiment, J. Statist. plann. Inference, 5 (1981), 381–389.
S. Ghosh, Influential nonnegligible parameters under the search linear model, Commun. Statist. Theory Methods., 16 (1987), 1013–1025.
S. Ghosh, Sequential assembly of fractions in factorial experiments, in: Design and Analysis of Experiments: Handbook of Statistics 13 (ed: S. Ghosh and C. R. Rao), Elsevier (Amsterdam, 1996), 407–435.
S. Ghosh and C. Burns, Comparison of four new general classes of search designs, Australian & New Zealand J. of Statist., 44(3) (2002), 357–366.
S. Ghosh and C. R. Rao, Design and Analysis of Experiments, North-Holland, Elsevier Science B.V. (Amsterdam, 1996).
S. Ghosh and H. Talebi, Main effect plans with additional search property for 2m factorial experiments, J. Statist. Plann. Inference, 36 (1993), 367–384.
S. Ghosh and T. Teschmacher, Comparisons of search designs using search probabilities, J. Statist. Plann. Inference, 104 (2002), 439–458.
B. C. Gupta, A bound connected with 26 factorial search designs of resolution 3.1., Commu. Statist.-Theor. Meth., 17(9) (1988), 3137–3144.
B. C. Gupta and S. S. R. Carvajal, A necessary condition for the existence of main effect plus one plans for 2m factorials, Commu. Statist.-Theor. Meth., 13(5) (1984), 567–580.
G. Katona and J. N. Srivastava, Minimal 2-coverings of a finite affine space based on GF(2), J. Statist. Plann. Inference, 8 (1983), 375–388.
R. Mukerjee and K. Chatterjee, An application of Hadamard matrices for the construction of main effect plus two plans for 2m factorials, Utilitas Math., 45 (1994), 213–218.
T. Ohnishi and T. Shirakura, Search designs for 2m factorial experiments, J. Statist. Plann. Inference, 11 (1985), 241–245.
R. L. Plackett and J. P. Burman, The design of optimum multifactorial experiments, Biometrika, 33 (1946), 305–325.
T. Sawa, Information criteria for discriminating among alternative regression models, Econometrica, 46 (1978), 1273–1282.
T. Shirakura, Main effect plus one or two plans for 2m factorials, J. Statist. Plann. Inference, 27 (1991), 65–74.
T. Shirakura, Fractional factorial designs of two and three levels, Discrete Math., 116 (1993), 99–135.
T. Shirakura and T. Ohnishi, Search designs for 2m factorials derived from balanced arrays of strength 2(ℓ + 1) and AD-optimal search designs, J. Statist. Plann. Inference., 11 (1985), 247–258.
T. Shirakura, T. Suetsugu, and T. Tsuji, Construction of main effect plus two plans for 2m factorials, J. Statist. Plann. Inference, 105 (2002), 405–415.
T. Shirakura, T. Takahashi and J. N. Srivastava, Search probabilities for nonzero effects in search designs for the noisy case, Ann. Statist., 24 (1996), 2560–2568.
T. Shirakura and S. Tazawa, Enumeration a representation of nonisomorphic weighted graphs, Kobe J. Math., 3 (1986), 229–235.
T. Shirakura and S. Tazawa, Series of main effect plus one or two plans for 2m factorilas when three-factor an higher order interactions are negligible, J. Japan Statist. Soc., 21 (1991), 211–219.
T. Shirakura and S. Tazawa, A series of search designs for 2m factorial designs of resolution V which permit search of one or two unknown extra three-factor interations, Ann. Inst. Statist. Math., 44 (1992), 185–196.
T. Shirakura and S. Tazawa, Search designs for balanced 2m factorial designs of resolution 2ℓ+1 when (ℓ+s+1)-factor and higher order interaction are negligible, J. Comb., Infor., Sys. Sci., 18 (1993), 233–246.
J. N. Srivastava, Some general existence conditions for balanced arrays of strength t and 2 symbols, J. Comb. Theory, 13 (1972), 198–206.
J. N. Srivastava, Application of the information function to dimensionality analysis and curved manifold clustering in: Proceedings of the third International Symposium on Multivariate Analysis (ed. P. R. Krishnaiah), Academic Press (New York, 1973), 369–381.
J. N. Srivastava, Designs for searching non-negligible effects, in: A Survey of Statistical Design and Linear Models (ed. J. N. Srivastava), North-Holland, Elsevier Science B.V. (Amsterdam, 1975), 507–519.
J. N. Srivastava, Optimal search design, or designs optimal under bias free optimality criteria, in: Statistical Decision Theory and Related Topics II (eds. S. S. Gupta and D. S. Moore), Academic Press (New York, 1977), 375–409.
J. N. Srivastava, Sensitivity and revealing power: Two fundamental statistical criteria other than optimality arising in discrete experimentation, in: Experimental Designs, Statiscical Models, and Genetic Statistics Models, and Genetic Statistics (ed. K. Hinkelmann), Marcel Dekker (New York, 1984), 95–117.
J. N. Srivastava, A 28 factorial main effect plus plan with good revealing power, Sankhya, 54 (1992), 461–474.
J. N. Srivastava, Nonadditivity in row-column designs, J. Combin, Inf. Sys., 18, 85–96.
J. N. Srivastava, A critique of some aspects of experimental design, in: Design and Analysis of Experiments: Handbook of Statistics 13 (eds. S. Ghosh and C. R. Rao), Elsevier (Amsterdam, 1996), 309–341.
J. N. Srivastava and S. Arora, An infinite series of resolution III.2 designs for the 2m factorial experiment, Discrete Math., 98 (1991), 35–56.
J. N. Srivastava and R. J. Beaver, On the superiority of the nested multidimensional block designs, relative to the classical incomplete block designs, J. Statist. Plann. Inference, 13 (1986), 133–150.
J. N. Srivastava and J. Y. Chu, Multistage design procedures for identifying two factor interactions, when higher effects are negligible, J. Statist. Plann. Inference, 78 (1999), 149–189.
J. N. Srivastava and S. Ghosh, A series of balanced 2m factorial designs of resolution V which allow search and estimation of one extra unknown effect, Sankhya (Ser. B), 38 (1976), 280–289.
J. N. Srivastava and S. Ghosh,Balanced 2m factorial designs of resolution V which allow search and estimation of one extra unknown effect, 4 <-m <-8, Commu. Statist.-Theor. Meth., 6 (1977), 141–166.
J. N. Srivastava and S. Ghosh, On non-orthogonality and non-optimality of Addelman’s main effect plans satisfying the condition of proportional frequencies, Statistics and Probability Letters., 26 (1996), 51–60.
J. N. Srivastava and B. C. Gupta, Main effect plan for 2m factorials which allow search and estimation of one unknown effect, J. Statist. Plann. Inference, 3 (1979), 259–265.
J. N. Srivastava and Y. C. Wang, Row-column designs: non-additivity makes them hazardous to use, J. Statist. Plann. Inference, 73 (1998), 277–315.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2007 János Bolyai Mathematical Society and Springer-Verlag
About this chapter
Cite this chapter
Ghosh, S., Shirakura, T., Srivastava, J.N. (2007). Model Identification Using Search Linear Models and Search Designs. In: Csiszár, I., Katona, G.O.H., Tardos, G., Wiener, G. (eds) Entropy, Search, Complexity. Bolyai Society Mathematical Studies, vol 16. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-32777-6_4
Download citation
DOI: https://doi.org/10.1007/978-3-540-32777-6_4
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-32573-4
Online ISBN: 978-3-540-32777-6
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)