理論上言之，貝氏估計式是在貝氏風險下的最佳解，本文不就此論點做驗證。
反之我們想研究若以在平方誤損失函數下所得之風險函數為比較準則，各類估計式的表現。
由模擬結果顯示，各估計式之表現會隨著參數值範圍的不同而改變，但整體言之，只要樣本數夠大，最大概似估計式仍是最好的選擇。

1.簡介
2.三參數廣義貝它分佈
3.最大概似估計式
4.貝氏估計式
5.模擬研究
6.結論

[1] Abramowitz, M. and I. A. Stegun (1965), Handbook of Mathematical Functions, Dover Publications, Sections 6.3 and 6.4。
[2] J. J. Chen and M. R. Novick (1984),``Bayesian analysis for binomial models with generalized beta prior distribution", emph{Journal of Educational Statistics}, 9, 163-175。
[3] Bonnie Kegan and Webster West (2005), ``Modeling the simple epidemic with deterministic differential equations and random initial conditions", emph{Mathematical Biosciences}, 195, 179-193。
[4] D. L. Libby (1981), ``Multivariate fixed state utility assemessment", PhD dissertation, University of Iowa, Iowa City。
[5] D. L. Libby and M. R. Novick (1982), ``Multivariate generalized beta distributions with applications to utility assessment", emph{Journal of Educational Statistics}, 7(4), 271-294。
[6] E. L. Lehmann (1983), Theory of point estimation, New York Wiley。
[7] D. V. Lindley (1980), ``Approximate Bayesian methods", emph{Trabajos Estadistica}, 31, 223-237。
[8] Christian P. Robert and George Casella (2004), Monte Carlo Statistical Methods, second edition, Springer。
[9] M. Sahinoglu (1981), ``Statistical inference on the reliability performance index for electric power generation systems", PhD dissertation, Institute of statistics, College Station, Texas A&M University。
[9] M. Sahinoglu and D. L. Libby and S. R. Das (2005), ``Measuring Availability Indices with Small Samples for Component and Network Reliability using the Sahinoglu-Libby Probability Model", emph{IEEE, Instrumentation and Measurement}, 54(3), June。
[10] A. Zellner (1986), ``Bayesian estimation and prediction using asymmetric loss function", emph{Journal of the American Statistical Association}, 81(394), 446-451。