在迴歸分析中，離群值與不相關變數常會影響模型估計、變數選取與預測表現。懲罰加權最小絕對偏差 LASSO（Penalized Weighted Least Absolute Deviation LASSO，PWLAD-LASSO）結合觀測值加權與自適應 LASSO（Adaptive LASSO）懲罰項，提供一個同時處理穩健估計、離群值偵測與變數選取的架構。然而，當資料中存在群集型污染，或部分不相關變數在解釋變數空間中形成誤導性的群集結構時，PWLAD-LASSO 的初始權重建構可能受到干擾，進而影響後續模型表現。本研究提出一個改良式 PWLAD-LASSO 方法，透過前向變數篩選程序（forward variable screening procedure），在模型估計前先辨識可能造成誤導性群集結構的變數。此方法利用變數對資料群集結構的影響程度衡量其不穩定性，並透過模型比較準則檢查候選變數是否應重新加入模型，使有助於模型配適的變數能被保留，而主要干擾初始權重建構的變數則可被移除。本研究設計不同污染型態的模擬實驗，評估所提出方法在離群值偵測、變數選取、係數估計與預測表現上的影響，並進一步使用不動產估價資料集比較實際資料上的預測能力。研究結果顯示，所提出的篩選程序能降低誤導性群集結構對 PWLAD-LASSO 的影響，並提升模型在污染資料下的穩定性與預測表現。

In regression analysis, outliers and irrelevant variables often affect model estimation, variable selection, and prediction performance. The Penalized Weighted Least Absolute Deviation LASSO (PWLAD-LASSO) combines observation weighting with an adaptive LASSO penalty, providing a framework for robust estimation, outlier detection, and variable selection. However, when the data contain clustered contamination or irrelevant variables that induce misleading clustering structures, the initial weight construction of PWLAD-LASSO may be disturbed. This study proposes an improved PWLAD-LASSO approach based on a forward variable screening procedure. The proposed procedure identifies variables that may induce misleading clustering structures before model estimation. It measures the instability of each variable according to its influence on the clustering structure of the data and uses a model comparison criterion to determine whether candidate variables should be retained. Simulation studies under different contamination settings and decoy-variable structures are conducted to evaluate the proposed method in terms of outlier detection, variable selection, coefficient estimation, and prediction performance. A real data analysis using the Real Estate Valuation data set is also performed to compare predictive accuracy. The results show that the proposed screening procedure reduces the influence of misleading clustering structures and improves the stability and prediction performance of PWLAD-LASSO.

Alfons, A., Croux, C., and Gelper, S. (2013). Sparse least trimmed squares regression for analyzing high-dimensional large data sets. The Annals of Applied Statistics, 7(1):226–248.
Arslan, O. (2012). Weighted lad-lasso method for robust parameter estimation and variable selection in regression. Computational Statistics & Data Analysis, 56(6):1952–1965.
Fan, J., Fan, Y., and Barut, E. (2014). Adaptive robust variable selection. The Annals of Statistics, 42(1):324–351.
Fan, J. and Li, R. (2001). Variable selection via nonconcave penalized likelihood and its oracle properties. Journal of the American Statistical Association, 96(456):1348–1360.
Giloni, A. and Padberg, M. (2006). The theory and practice of linear programming for lad regression. Computational Statistics & Data Analysis.
Hampel, F. R., Ronchetti, E. M., Rousseeuw, P. J., and Stahel, W. A. (1986). Robust Statistics:The Approach Based on Influence Functions. Wiley.
Hardin, J. and Rocke, D. M. (2004). Outlier detection in the multiple cluster setting using the minimum covariance determinant estimator. Computational Statistics & Data Analysis,44(4):625–638.
Hastie, T., Tibshirani, R., and Friedman, J. (2009). The Elements of Statistical Learning. Springer.
Huang, J., Ma, S., and Zhang, C.-H. (2008). Adaptive lasso for sparse high-dimensional regression models. Statistica Sinica, 18(4):1603–1618.
Huber, P. J. (1981). Robust Statistics. Wiley.
Jiang, Y., Wang, Y., Zhang, J., Xie, B., Liao, J., and Liao, W. (2021). Outlier detection and robust variable selection via the penalized weighted lad-lasso method. Journal of Applied Statistics, 48(2):234–246.
Koenker, R. (2005). Quantile Regression. Cambridge University Press.
Koenker, R. and Bassett, G. (1978). Regression quantiles. Econometrica, 46(1):33–50.
Kosinski, A. S. (1998). A procedure for the detection of multivariate outliers. Computational Statistics & Data Analysis, 29(2):145–161.
Maronna, R. A., Martin, R. D., Yohai, V. J., and Salibian-Barrera, M. (2019). Robust Statistics:Theory and Methods (with R). John Wiley & Sons, Hoboken, NJ, 2 edition.
Narula, S. C. and Wellington, J. F. (1982). The minimum sum of absolute errors regression: A state of the art survey. International Statistical Review / Revue Internationale de Statistique, 50(3):317–326.
Rousseeuw, P. J. and Leroy, A. M. (1987). Robust Regression and Outlier Detection. Wiley.
Smucler, E. and Yohai, V. J. (2017). Robust and sparse estimators for linear regression models. Computational Statistics & Data Analysis, 111:116–130.
Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B, 58(1):267–288.
Wang, L., Li, H., and Leng, C. (2007). Robust regression shrinkage and consistent variable selection through the lad-lasso. Journal of Business & Economic Statistics, 25(3):347–355.
Wang, X., Jiang, Y., Huang, M., and Zhang, H. (2013). Robust variable selection with exponential squared loss. Journal of the American Statistical Association, 108(502):632–643.
Xu, J. and Ying, Z. (2010). Simultaneous estimation and variable selection in median regression using lasso-type penalty. Annals of the Institute of Statistical Mathematics, 62(3):487–514.
Yeh, I.-C. and Hsu, T.-K. (2018). Building real estate valuation models with comparative approach through case-based reasoning. Applied Soft Computing, 65:260–271.
Zhao, P. and Yu, B. (2006). On model selection consistency of lasso. Journal of Machine Learning Research, 7:2541–2563.
Zou, H. (2006). The adaptive lasso and its oracle properties. Journal of the American Statistical Association, 101(476):1418–1429.
Zou, H. and Hastie, T. (2005). Regularization and variable selection via the elastic net. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 67(2):301–320.