题名

二元數列之預測方法的模擬比較:預測方法是否需具隨機性機制?

并列篇名

Simulated Comparison on Some Predictors for Binary Sequences: To Randomize or Not to Randomize?

DOI

10.29973/JCSA.200606.0006

作者

蘇靜蓮(Jing-Lien Su);曾玉玲(Yu-Ling Tseng)

关键词

二元數列 ; 預測方法 ; 隨機性 ; 多數決 ; 馬可夫鏈 ; 大樣本理論 ; 柏努利 ; predictor ; binary sequence ; prediction algorithm ; randomization ; play-the-winner ; asymptotic property ; Markov chain ; Bernoulli

期刊名称

中國統計學報

卷期/出版年月

44卷2期(2006 / 06 / 01)

页次

198 - 221

内容语文

繁體中文

中文摘要

在本文中,首先我們想要比較不同預測方法在有限樣本時的優劣情形,並且想要知道預測方法中的隨機性機制是否對於預測二元數列有實質的幫助。透過模擬研究結果發現,對於幾乎所有的二元數列之預測問題而言,任何類型的隨機性對於預測是無益的,因此建議使用多數決方法來預測二元數列。另外,我們亦證明一些大樣本理論的結果,特別是當預測由馬可夫鏈模型產生的數列時,我們提供一預測方法要具漸近最佳性質之充分條件的理論刻畫。

英文摘要

In this work, we compare the finite sample performances of some prediction algorithms through simulation studies. We also want to see if the randomization is really beneficial in predicting binary sequences. Our study indicates that any sort of randomization is not really favorable for predicting almost all kind of binary data, hence the play-the-winner strategy is suggested. Our study also establishes some asymptotic results, in particular we characterize a sufficient condition for a prediction rule to be asymptotically optimal under the Markov chain scheme.

主题分类 基礎與應用科學 > 統計
参考文献
  1. Feder, M.,Merhav, N.,Gutman, M.(1992).Universal Prediction of individual Sequences.IEEE Trans. Inform. Theory,38,1258-1270.
  2. Gupta, S. S.,Berger, J. O.(1994).Statistical Decision Theory and Related Topics V.Berlin:Springer.
  3. Hutter, M.(2001).New Error Bounds for Solomonoff Prediction.J. Comput. Syst. Sci.,62(4),653-667.
  4. Lerche, H. R.,Sarkar, J.(1994).Statistical Decision Theory and Related Topics V.Berlin:Springer.
  5. Nastaj, J.,Ambrozek, B.(2001).Binary Adsorption Capacity Prediction of Water and Water Immiscible Organic Compounds on Activated Carbon.Inz. Chem. Procesowa,22(3D),995-1000.
  6. Ross, S. M.(1997).Introduction to Probability Models.Academic Press.
  7. Su, J. L.(2002).Hualien,Department of Applied Mathematics, National Dong Hwa University.