A Comparative analysis of multiple outlier detection procedures in the linear regression model

• Published in: Computational statistics & data analysis Vol. 36; no. 3; pp. 351 - 382
• Main Authors: Wisnowski, James W, Montgomery, Douglas C, Simpson, James R
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 28.05.2001; Elsevier Science; Elsevier
• Series: Computational Statistics & Data Analysis
• Subjects: Exact sciences and technology; Linear inference, regression; Mathematics; Minimum volume ellipsoid; Monte Carlo simulation; Multiple outliers; Outlier; Probability and statistics; Robust regression; Sciences and techniques of general use; Statistics; Outlier; Minimum volume ellipsoid; Robust regression; Monte Carlo simulation; Multiple outliers; Monte Carlo method; Linear regression; Statistical estimation; Regression residual; Algorithm; Linear model; Multiple regression; Direct method; Statistical regression; Ordinary least squares residual; Least median od squares estimator; Simulation; Regression model; Numerical simulation; Robustness; Linear estimation; Distance; M estimation
• ISSN: 0167-9473; 1872-7352
• DOI: 10.1016/S0167-9473(00)00042-6

We evaluate several published techniques to detect multiple outliers in linear regression using an extensive Monte Carlo simulation. These procedures include both direct methods from algorithms and indirect methods from robust regression estimators. We evaluate the impact of outlier density and geometry, regressor variable dimension, and outlying distance in both leverage and residual on detection capability and false alarm (swamping) probability. The simulation scenarios focus on outlier configurations likely to be encountered in practice and use a designed experiment approach. The results for each scenario provide insight and limitations to performance for each technique. Finally, we summarize each procedure's performance and make recommendations.