Robust statistics for outlier detection

By Peter J. Rousseeuw, Mia Hubert. 2011

810 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

BoostClean: Automated Error Detection and Repair for Machine Learning (2017, methodology): “A recently popular approach is a robust estimator of the sample population variance called Minimum Covariance Determinant [44] (MCD), which has been used in systems such as MacroBase [9].”
Results of the WMT20 Metrics Shared Task (2020, methodology): “This method, recommended by the statistics literature (Iglewicz and Hoaglin, 1993; Rousseeuw and Hubert, 2011; Leys et al., 2013) depends on the median and the median absolute deviation (MAD) which is the median of the absolute difference between each point and the median.”
Unsupervised Energy-based Out-of-distribution Detection using Stiefel-Restricted Kernel Machine (2021, background): “Several approaches to anomaly detection exist, such as in the domain of clustering and robust statistics (for a comprehensive review, see [3], [4]).”
DataComp: In search of the next generation of multimodal datasets (2023, background): “Classical work considers dataset cleaning and outlier removal [65, 140, 115, 116] to discard samples that may lead to undesirable model bias.”
Tangled up in BLEU: Reevaluating the Evaluation of Automatic Machine Translation Evaluation Metrics (2020, methodology): “Instead, we use the median and the Median Absolute Deviation (MAD) which are more robust (Iglewicz and Hoaglin, 1993; Rousseeuw and Hubert, 2011; Leys et al., 2013).”
Generalized Out-of-Distribution Detection: A Survey (2021, neutral): “Although these methods suffer from the curse of dimensionality on images, they can be alleviated by dimensionality reduction methods [280], [281] and the NN-based density methods (c.”
Progress in Outlier Detection Techniques: A Survey (2019, background): “Rousseeuw and Hubert [117] also gave a broader discussion of statistical techniques for outlier detection.”
PointCleanNet: Learning to Denoise and Remove Outliers from Dense Point Clouds (2019, background): “…classical approaches for outlier detection, classiﬁcation and removal have been proposed primarily in the statistics and data mining communities, in the general setting of point clouds in arbitrary dimensions, with several monographs dedicated speciﬁcally to this topic [Pin95,BG05,RH11,Agg15].”
Explaining Extreme Events of 2015 from a Climate Perspective (2016, methodology): “The ellipse denotes a quantile of 5% multivariate extremes computed by Hotelling’s T^2 control chart (Santos-Fernández 2012) using robust mean and covariance estimates (Rousseeuw and Hubert 2011).”
Actively Semi-Supervised Deep Rule-based Classifier Applied to Adverse Driving Scenarios (2019, background): “According to [27], the mσ rule states that nearly all (99.”
Novelty Detection and Learning from Extremely Weak Supervision (2019, methodology): “The xClass exploratory mechanism is able to discover new classes as they arrive to the system due to its mechanism based on the recursive density estimation [19] and Chebyshev inequality approach [22] as given by Fig.”