共變異數反矩陣的估計

Estimation of covariance inverse

余政遠(Zheng-Yunan Yu)

指導教授 ： 許文郁

國立清華大學/理學院/統計學研究所/碩士(2013年)

https://doi.org/10.6843/NTHU.2013.00624

摘要

針對共變異數反矩陣在高維度情況下的估計，近年來所提出的方式都是對於配適程度與模型簡化中間做一個折衷的方式，採用L1懲罰項的最大概似估計量，但是儘管L1的限制式能有效的找出稀疏的信號，卻會導致估計量有被壓縮的情況產生；為了改善這個問題，我們將找出稀疏的信號與估計這兩個步驟分開進行，希望能有效降低壓縮所導致的誤差。

關鍵字

共變異數反矩陣

並列摘要

In recent years ,the estimate of covariance inverse matrix with high dimemsional which between model fitting and model simplifies .Many scholars have suggested that using maximum penalized likelihood function with L1-norm .The L1-norm can identify sparse signal effectively ,but lead to estimator be compressed .To improve this problem ,we will find the location of the spares signal and estimate these separately ,hoping to reduce the resulting of compression of the estimator .

並列關鍵字

covariance inverse

參考文獻

1. N. Meinshausen and P. Buhlmann, (2006). High dimensional graphs and variable selection with the lasso. Annals of statistics, 34:1436-1462

minimization approach to sparse precision matrix estimation. J. American Statistical Association 106, 594-607.

4. Cai, T. & Liu, W. (2011). A direct estimation approach to sparse linear discriminant analysis. J. American Statistical Association 106, 1566-1577.

5. J. Friedman, T. Hastie, and R. Tibshirani. (2010). Regularization paths for generalized linear models via coordinate descent. Journal of Statistical Software, 33(1):1,.

6. Vandenber, L., Boyd, S. & Wu, S.-P. (1998). Determinant maximization with linear matrix inequality constraints. SIAM J. Matrix Anal. Appl. 19, 499–533.