AIC、BIC和EBIC之回顧

Review of AIC, BIC and EBIC

顏妤樺(Yu-Hua Yen)

指導教授 ： 洪慧念

國立交通大學/理學院/統計學研究所/碩士(2014年)

https://doi.org/10.6842/NCTU.2014.00221

摘要

自資訊爆炸以來，利用統計方法分析資料漸漸成為一種常態。而我們所面對的問題也從過去的大樣本資料分析逐漸轉變成高維度資料分析。如何找出這些資料的最適模型是我們最重要的課題。在這篇文章中，我們將Chen & Chen (2008)提出之針對高維度模型選取方法EBIC與常見的模型選取方法AIC、BIC做比較，並利用模擬的方式說明這些方法的差異與優劣。

關鍵字

高維度模型 ； 模型選取 ； AIC ； BIC ； EBIC

並列摘要

Since the information explosion, analyzing data by using statistical methods progressively becomes norm. Nowadays, the problem we are faced with large sample size analysis gradually transformed into high dimensional model analysis. How to find the optimal model for the data is our most important issue. In our study, we compare EBIC, which proposed by Chen & Chen (2008) for high dimensional model, with common model selection methods, AIC and BIC, and use simulations illustrating the difference and the pros and cons of these methods.

並列關鍵字

High Dimensional Model ； Model Selection ； AIC ； BIC ； EBIC

參考文獻

[1] Akaike, H. (1974), “A New Look at the Statistical Model Identification,” IEEE Transactions on Automatic Control, 19, 716-723.

[2] Burnham, K. P. and Anderson, D. R. (2004), “Multimodel Inference: Understanding AIC and BIC in Model Selection,” Sociological Method & Research, 33, 261-304.

[3] Chen, J. and Chen, Z. (2008), “Extended Bayesian information criteria for model selection with large model spaces,” Biometrika, 95, 759-771.

[4] Edwards, D., Abreu, G. C. G. and Labouriau, R. (2010), “Selecting high-dimensional mixed graphical models using minimal AIC or BIC forests,” Edwards et al. BMC Bioinformatics, 11:18.

[5] Konishi, S. and Kitagawa, G. (2008), Information Criteria and Statistical Modeling, New York: Springer.