Ensembles for unsupervised outlier detection

Zimek, Arthur; Campello, Ricardo J. G. B.; Sander, Jöerg

doi:10.1145/2594473.2594476

Cited by 219 publications

(90 citation statements)

References 73 publications

(61 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For better comparability of the algorithms' outputs, we rank them by computing the percentiles of the algorithm scores. These are then aggregated into ensemble scores by computing the minimum (consensus voting), the mean (balanced voting), or the maximum (risky voting) of the scores of selected well-performing algorithms (e.g., Aggarwal, 2012;Zimek et al, 2013).…”

Section: Ensembles Of Anomaly Detection Algorithmsmentioning

confidence: 99%

“…The selection of algorithms for computing the ensemble is a compromise between accurate detection of and diversity amongst the selected algorithms (Zimek et al, 2013). We select the four best algorithms (KDE, REC, KNN-Gamma, T 2; referred to together as 4b) and the three best distance-based algorithms (KDE, REC, KNN-Gamma; referred to together as 3d) for computing their ensembles.…”

Section: Ensemblesmentioning

confidence: 99%

“…The superior performance of the maximum choice compared to the minimum indicates that single algorithms overlook more often anomalous events than raising false alarm. Nevertheless, the maximum has the caveat that even a single algorithm may cause a false alarm (Zimek et al, 2013), e.g., due to a poor parameterization or inadequate assumptions about properties of the data. Thus, a more balanced voting procedure like the mean is the preferable choice and more stable with respect to possible error sources.…”

Section: Ensemblesmentioning

confidence: 99%

See 2 more Smart Citations

Multivariate anomaly detection for Earth observations: a comparison of algorithms and feature extraction techniques

et al. 2017

View full text Add to dashboard Cite

Abstract. Today, many processes at the Earth's surface are constantly monitored by multiple data streams. These observations have become central to advancing our understanding of vegetation dynamics in response to climate or land use change. Another set of important applications is monitoring effects of extreme climatic events, other disturbances such as fires, or abrupt land transitions. One important methodological question is how to reliably detect anomalies in an automated and generic way within multivariate data streams, which typically vary seasonally and are interconnected across variables. Although many algorithms have been proposed for detecting anomalies in multivariate data, only a few have been investigated in the context of Earth system science applications. In this study, we systematically combine and compare feature extraction and anomaly detection algorithms for detecting anomalous events. Our aim is to identify suitable workflows for automatically detecting anomalous patterns in multivariate Earth system data streams. We rely on artificial data that mimic typical properties and anomalies in multivariate spatiotemporal Earth observations like sudden changes in basic characteristics of time series such as the sample mean, the variance, changes in the cycle amplitude, and trends. This artificial experiment is needed as there is no "gold standard" for the identification of anomalies in real Earth observations. Our results show that a well-chosen feature extraction step (e.g., subtracting seasonal cycles, or dimensionality reduction) is more important than the choice of a particular anomaly detection algorithm. Nevertheless, we identify three detection algorithms (k-nearest neighbors mean distance, kernel density estimation, a recurrence approach) and their combinations (ensembles) that outperform other multivariate approaches as well as univariate extreme-event detection methods. Our results therefore provide an effective workflow to automatically detect anomalies in Earth system science data.

show abstract

Section: Ensembles Of Anomaly Detection Algorithmsmentioning

confidence: 99%

Section: Ensemblesmentioning

confidence: 99%

Section: Ensemblesmentioning

confidence: 99%

See 1 more Smart Citation

Multivariate anomaly detection for Earth observations: a comparison of algorithms and feature extraction techniques

et al. 2017

View full text Add to dashboard Cite

show abstract

“…Contrasting to these work which aims to improve the scalability in traditional k NN methods via subsampling, ZERO++ utilises the probability of zero appearances to introduce a new anomaly scoring measure. Also, they operate in a single subsample only and use the full dimensionality to compute the kth-NN distance, which can lead to unstable detection performance and have the curse of dimensionality problem in high-dimensional data (Zimek, Campello, & Sander, 2013a), while ZERO++ builds a set of models in a set of subsamples and subspaces.…”

Section: Methods For Numeric Datamentioning

confidence: 99%

ZERO++: Harnessing the Power of Zero Appearances to Detect Anomalies in Large-Scale Data Sets

Pang¹,

Ting²,

Albrecht³

et al. 2016

jair

View full text Add to dashboard Cite

This paper introduces a new unsupervised anomaly detector called ZERO++ which employs the number of zero appearances in subspaces to detect anomalies in categorical data. It is unique in that it works in regions of subspaces that are not occupied by data; whereas existing methods work in regions occupied by data. ZERO++ examines only a small number of low dimensional subspaces to successfully identify anomalies. Unlike existing frequencybased algorithms, ZERO++ does not involve subspace pattern searching. We show that ZERO++ is better than or comparable with the state-of-the-art anomaly detection methods over a wide range of real-world categorical and numeric data sets; and it is efficient with linear time complexity and constant space complexity which make it a suitable candidate for large-scale data sets.

show abstract

“…These issues are interrelated and refer to the problems of model selection and assessment (evaluation or validation) of results in unsupervised learning. These problems have been investigated for decades in the area of unsupervised data clustering [14], but are rarely mentioned and are virtually untouched in the area of outlier detection [43].…”

Section: Ssdbm '15mentioning

confidence: 99%

On the internal evaluation of unsupervised outlier detection

Marques

Campello

Zimek

2015

Proceedings of the 27th International Conference on Scientific and Statistical Database Management

Self Cite

View full text Add to dashboard Cite

Although there is a large and growing literature that tackles the unsupervised outlier detection problem, the unsupervised evaluation of outlier detection results is still virtually untouched in the literature. The so-called internal evaluation, based solely on the data and the assessed solutions themselves, is required if one wants to statistically validate (in absolute terms) or just compare (in relative terms) the solutions provided by different algorithms or by different parameterizations of a given algorithm in the absence of labeled data. However, in contrast to unsupervised cluster analysis, where indexes for internal evaluation and validation of clustering solutions have been conceived and shown to be very useful, in the outlier detection domain this problem has been notably overlooked. Here we discuss this problem and provide a solution for the internal evaluation of top-n (binary) outlier detection results. Specifically, we propose an index called IREOS (Internal, Relative Evaluation of Outlier Solutions) that can evaluate and compare different candidate labelings of a collection of multivariate observations in terms of outliers and inliers. We also statistically adjust IREOS for chance and extensively evaluate it in several experiments involving different collections of synthetic and real data sets.

show abstract

Ensembles for unsupervised outlier detection

Cited by 219 publications

References 73 publications

Multivariate anomaly detection for Earth observations: a comparison of algorithms and feature extraction techniques

Multivariate anomaly detection for Earth observations: a comparison of algorithms and feature extraction techniques

ZERO++: Harnessing the Power of Zero Appearances to Detect Anomalies in Large-Scale Data Sets

On the internal evaluation of unsupervised outlier detection

Contact Info

Product

Resources

About