论文信息 - Finding multivariate outliers with FastPCS

Finding multivariate outliers with FastPCS

The Projection Congruent Subset (PCS) is a new method for finding multivariate outliers. Like many other outlier detection procedures, PCS searches for a subset which minimizes a criterion. The difference is that the new criterion was designed to be insensitive to the outliers. PCS is supported by FastPCS, a fast and affine equivariant algorithm which is also detailed. Both an extensive simulation study and a real data application from the field of engineering show that FastPCS performs better than its competitors.

Eric Schmitt | Kaveh Vakili

[1] Deepayan Sarkar,et al. Lattice: Multivariate Data Visualization with R , 2008 .

[2] Peter J. Rousseeuw,et al. Unconventional features of positive-breakdown estimators , 1994 .

[3] P. Rousseeuw,et al. Unmasking Multivariate Outliers and Leverage Points , 1990 .

[4] D. Berry,et al. Statistics: Theory and Methods , 1990 .

[5] R Core Team,et al. R: A language and environment for statistical computing. , 2014 .

[6] V. Yohai,et al. Robust Statistics: Theory and Methods , 2006 .

[7] Gabor Grothendieck,et al. Lattice: Multivariate Data Visualization with R , 2008 .

[8] Yohai Victor,et al. The maximum bias of robust covariances , 1990 .

[9] I-Cheng Yeh,et al. Modeling slump flow of concrete using second-order regressions and artificial neural networks , 2007 .

[10] Peter J. Rousseeuw,et al. Robust regression and outlier detection , 1987 .

[11] Katrien van Driessen,et al. A Fast Algorithm for the Minimum Covariance Determinant Estimator , 1999, Technometrics.

[12] Peter Filzmoser,et al. An Object-Oriented Framework for Robust Multivariate Analysis , 2009 .

[13] David L. Woodruff,et al. Identification of Outliers in Multivariate Data , 1996 .