摘要(英) |
Online testing or multivariate experiment is a method that grows rapidly
in the digital world for various applications, such as website and email
optimization. Due to the highly development of technology, online testings are more often conducted on other platforms, such as laptops, smartphones, and smart watches. Thus, designing efficient online testings across
multiple platforms is an urgent issue. Sadeghi et al. (2016) proposed
a design selection criterion for two-level multiple-platform experiments.
Sadeghi et al. (2017) further extended Sadeghi et al. (2016) to fourlevel factors. However, such naive modifications lack strong statistical justifications. In this thesis, we study such design problem through two approaches, referred to as Bayesian method and Taylor Series method. We
also provide optimal designs based on our methods for various run sizes and
numbers of factors. |
參考文獻 |
References
[1] Sadeghi, S., Qian, P. Z. G., and Arora, N. ”Sliced designs for multi-platform
online experiments.” (2016).
[2] Sadeghi, S., Qian, P. Z. G., and Arora, N. ”Sliced minimum aberration designs
for four-platform experiments.” (2017).
[3] Fries, A., and Hunter, W. G. ”Minimum aberration 2 k–p designs.” Technometrics 22.4 (1980): 601-608.
[4] Joseph, V. R. ”A Bayesian approach to the design and analysis of fractionated
experiments.” Technometrics 48.2 (2006): 219-229.
[5] Kang, L., and Joseph, V. R. ”Bayesian optimal single arrays for robust parameter design.” Technometrics 51.3 (2009): 250-261.
[6] Joseph, V. R., and Delaney, J. D. ”Functionally induced priors for the analysis
of experiments.” Technometrics 49.1 (2007): 1-11.
[7] Wu, C. F. J., and Hamada, M. S. Experiments: planning, analysis, and
optimization. Vol. 552. John Wiley Sons, 2011.
[8] Phoa, F. K. H. ”A Swarm Intelligence Based (SIB) method for optimization
in designs of experiments.” Natural Computing 16.4 (2017): 597-605.
[9] Eberhart, R., and Kennedy, J. ”A new optimizer using particle swarm theory.” MHS’95. Proceedings of the Sixth International Symposium on Micro
Machine and Human Science. IEEE, 1995. |