Angle-based outlier detection in high-dimensional data

By Hans-Peter Kriegel, Matthias Schubert, Arthur Zimek. Knowledge Discovery and Data Mining, 2008

934 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/0372728b9a2def008ef3240a62362f0afbfb5d43

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

Outlier Detection in Axis-Parallel Subspaces of High Dimensional Data (2009, background): “While LOF and ABOD are very competitive in lower dimensional data sets, their performance considerably deteriorates with higher dimensionality while SOD remains very stable at optimal values.”
HRN: A Holistic Approach to One Class Learning (2020, neutral): “…can be classiﬁed into the following categories: distance based methods [42, 3, 28, 31, 60], density based methods [38, 56, 6], mixture models [3, 43], one-class classiﬁcation based methods [75, 78, 39], deep learning based representation learning using auto-encoders [10, 89, 69, 7, 94] and…”
Identifying Multi-instance Outliers (2010, neutral): “There are a large number of outlier detection studies in previous literature [10, 11, 12, 15].”
MAD-GAN: Multivariate Anomaly Detection for Time Series Data with Generative Adversarial Networks (2019, methodology): “For example, the Angle-Based Outlier Detection (ABOD) method [19] and Feature Bagging (FB) method [20] deal with data by taking variable correlations into consideration.”
Complete Random Forest Based Class Noise Filtering Learning for Improving the Generalizability of Classifiers (2019, background): “Moreover, most of them are distance basedmeasurements and bound to deteriorate due to the notorious “curse of dimensionality” [39].”
LoOP: local outlier probabilities (2009, methodology): “ABOD is also quite stable, but at a lower level than LoOP or LDOF.”
HiCS: High Contrast Subspaces for Density-Based Outlier Ranking (2012, background): “The implementation details are given in Section IV-A.”
ECOD: Unsupervised Outlier Detection Using Empirical Cumulative Distribution Functions (2022, background): “Speciﬁcally, the 11 competitors are Angle-Based Outlier Detection (ABOD) [57], Clustering-Based Local Outlier Factor (CBLOF) [40], Histogram-based Outlier Score (HBOS) [30], Isolation Forest (IForest) [17], k Nearest Neighbors (KNN) [15], Lightweight On-line Detector of Anomalies (LODA) [58], Local…”
A survey on unsupervised outlier detection in high‐dimensional numerical data (2012, methodology): “…that an outlier detection method tailored for high-dimensional data was based on the use of angles instead of plain distances: The outlier detection method ABOD (angle-based outlier detection) [44] assesses the variance in angles between an outlier candidate and all other pairs of points.”
A Review on Soft Sensors for Monitoring, Control, and Optimization of Industrial Processes (2021, background): “[27], local outlier factor [28] and angle-based outlier factor [29].”
Hierarchical Density Estimates for Data Clustering, Visualization, and Outlier Detection (2015, methodology): “Recent global approaches base the decision not on Euclidean distances but on anglevariance (ABOD [Kriegel et al. 2008] and an efficient variant using random projections [Pham and Pagh 2012]).”
Interpreting and Unifying Outlier Scores (2011, background): “score ABOD [33] is not primarily based on conventional distance measures.”
Ensembles for unsupervised outlier detection: challenges and research questions a position paper (2014, methodology): “Some popular models are based on the distance between objects [37; 60; 4; 74], or on the density of the neighborhood of an object [9; 56; 34; 38; 42], or based on the variance of angles between object vectors [42; 58], or on other principles of outlierness in various domains [12; 13; 3].”
Loda: Lightweight on-line detector of anomalies (2016, methodology): “This phenomenon (collection of weak classifiers result in a strong classifier) is already a well established paradigm in supervised classification Kuncheva (2004); Freund et al (1996), but has not been demonstrated in unsupervised anomaly detection, as most ensemble systems used in anomaly detection”
Outlier Analysis (2013, methodology): “It is also noted that an angle-based outlier detection for high-dimensional data has been proposed in [269], though this method has been discussed in the chapter on extreme value analysis (Chapter 2), since this method is not a subspace exploration technique.”
A comparative evaluation of outlier detection algorithms: Experiments and analyses (2018, methodology): “As shown in [7], lof outperforms AngleBased Outlier Detection [16] and One-class SVM [26] when applied on real-world datasets for outlier detection, which makes it a good candidate for this benchmark.”
A New Local Distance-Based Outlier Detection Approach for Scattered Real-World Data (2009, background): “In real-world data repositories, it is hard to ﬁnd a dataset for evaluating outlier detection algorithms, because only for very few real-world datasets it is exactly known which objects are really behaving diﬀer-ently [KSZ08].”