Abstract
We propose randomized inference (RI), a new statistical inference approach. RI may be realized through a randomized estimate (RE) of a parameter vector, which is a random vector that takes values in the parameter space with a probability density function (PDF) that depends on the sample or sufficient statistics, such as the posterior distributions in Bayesian inference. Based on the PDF of an RE of an unknown parameter, we propose a framework for both the vertical density representation (VDR) test and the construction of a confidence region. This approach is explained with the aid of examples. For the equality hypothesis of multiple normal means without the condition of variance homogeneity, we present an exact VDR test, which is shown as an extension of one-way analysis of variance (ANOVA). In the case of two populations, the PDF of the Welch statistics is given using RE. Furthermore, through simulations, we show that the empirical distribution function, the approximated t, and the RE distribution function of Welch statistics are almost equal. The VDR test of the homogeneity of variance is shown to be more efficient than both the Bartlett test and the revised Bartlett test. Finally, we discuss the prospects of RI.
Article PDF
Similar content being viewed by others
Avoid common mistakes on your manuscript.
References
Efron B. R. A. Fisher in 21st century. Statist Sci, 1998, 13: 95–122
Fang K T, Yang Z H, Kotz S. Generation of multivariate distributions by vertical density representation. Statistics, 2001, 35: 281–293
Fisher R A. Inverse probability. Math Proc Cambridge Philos Soc, 1930, 26: 528–535
Fisher R A. The fiducial argument in statistical inference. Ann Euenics, 1935, 6: 391–398
Kendall M G, Stuart A. The Advanced Theory of Statistics. London: Charles Griffin, 1958
Lehmann E L. Fisher, Neyman-Person theories of testing hypothesis: One theory or two? J Amer Statist Assoc, 1993, 88: 1242–1249
Mao S S, Cheng Y M, Pu X L. Probability and Mathematical Statistics (in Chinese). Beijing: Higher Education Press, 2015
Martin R, Liu C. Inferential models: A framework for prior-free posterior probabilistic inference. J Amer Statist Assoc, 2013, 108: 301–313
Pang K W, Yang Z H, Hou S H, et al. Some further results of multivariate vertical density representation and its application. Statistics, 2001, 35: 463–477
Singh K, Xie M G, Strawderman W E. Confidence distribution (CD)-distribution estimator of a parameter. In: IMS Lecture Notes-Monograph Series Complex Datasets and Inverse Problems: Tomograph and Beyond, vol. 54. Hayward: Inst Math Statist, 2007, 132–150
Troutt M D. A theorem on the density of the density ordinate and an alternative derivation of the Box-Muler method. Statistics, 1991, 22: 436–466
Troutt M D, Pang W K, Hou S H. Vetical Density Representation and Its Application. Singapore: World Scientific, 2004
Xie M G, Liu R Y, Damaraju C V, et al. Incorporating external information in analysis of clinical trials with binary outcomes, confidence distribution and unifying framework for meta-analysis. Ann Appl Stat, 2013, 7: 342–368
Xie M G, Singh K. Confidence distribution, the frequentist distribution estimator of a parameter: A review. Internat Statist Rev, 2013, 81: 3–39
Xie M G, Singh K, Strawderman W E. Confidence distribution and unifying framework for meta-analysis. J Amer Statist Assoc, 2011, 106: 302–333
Yang Z H. Randomized Estimation and VDR Test (in Chinese). Beijing: Higher Education Press, 2013
Acknowledgements
The first author was supported by National Natural Science Foundation of China (Grant No. 11771032). The authors sincerely thank Professor Bocheng Wei, Shuyuan He, and Min-ge Xie for their support and helpful discussions. The authors also thank the anonymous referees for their comments, which have helped us to considerably improve the manuscript.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Zhang, Z., Dai, J. & Yang, Z. Randomized statistical inference: A unified statistical inference frame of frequentist, fiducial, and Bayesian inference. Sci. China Math. 63, 1007–1028 (2020). https://doi.org/10.1007/s11425-017-9325-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11425-017-9325-9