Anomaly detection in scientific data using joint statistical moments

DOI10.1016/J.JCP.2019.03.003MaRDI QIDQ2220587zbMATH OpenOpenAlexWikidataFDO

Authors Konduri Aditya, H. Kolla, W. Philip Kegelmeyer, Timothy M. Shead, Julia Ling, Warren L. Davis

Publication date 25 January 2021

Published in Journal of Computational Physics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1808.04498

zbMATH Keywords

scientific computinganomaly detectionHellinger distancetensor decompositionautoignitioncokurtosis

Mathematics Subject Classification ID

Multivariate distribution of statistics (62H10) Statistical turbulence modeling (76F55) Turbulent combustion; reactive turbulence (76F80)

Abstract: We propose an anomaly detection method for multi-variate scientific data based on analysis of high-order joint moments. Using kurtosis as a reliable measure of outliers, we suggest that principal kurtosis vectors, by analogy to principal component analysis (PCA) vectors, signify the principal directions along which outliers appear. The inception of an anomaly, then, manifests as a change in the principal values and vectors of kurtosis. Obtaining the principal kurtosis vectors requires decomposing a fourth order joint cumulant tensor for which we use a simple, computationally less expensive approach that involves performing a singular value decomposition (SVD) over the matricized tensor. We demonstrate the efficacy of this approach on synthetic data, and develop an algorithm to identify the occurrence of a spatial and/or temporal anomalous event in scientific phenomena. The algorithm decomposes the data into several spatial sub-domains and time steps to identify regions with such events. Feature moment metrics, based on the alignments of the principal kurtosis vectors, are computed at each sub-domain and time step for all features to quantify their relative importance towards the overall kurtosis in the data. Accordingly, spatial and temporal anomaly metrics for each sub-domain are proposed using the Hellinger distance of the feature moment metric distribution from a suitable nominal distribution. We apply the algorithm to a one-dimensional turbulent auto-ignition combustion case and demonstrate that the anomaly metrics reliably capture the occurrence of auto-ignition in relevant spatial sub-domains at the right time steps.

Recommendations

• Anomaly detection: a functional analysis perspective
• Sparse representation of multivariate extremes with applications to anomaly detection
• Anomaly detection based on an iterative local statistics approach
• Finding anomalous periodic time series
• ICS for multivariate outlier detection with application to quality control

Cites work

• scientific article; zbMATH DE number 1820149 (Why is no real title available?)
• A Multilinear Singular Value Decomposition
• Kurtosis as Peakedness, 1905–2014.R.I.P.
• Numerically stable, scalable formulas for parallel and online computation of higher-order multivariate central moments with arbitrary weights
• Scikit-learn: machine learning in Python
• Symmetric Tensors and Symmetric Tensor Rank
• Tensor Decompositions and Applications
• Tensor decompositions for learning latent variable models
• Three-dimensional direct numerical simulation of a turbulent lifted hydrogen jet flame in heated coflow: a chemical explosive mode analysis
• Trigger detection for adaptive scientific workflows using percentile sampling

Cited in

(2)

• Estimating Higher-Order Moments Using Symmetric Tensor Decomposition
• Scalable symmetric Tucker tensor decomposition

Describes a project that uses

Uses Software

• TensorToolbox
• Scikit
• LOF

This page was built for publication: Anomaly detection in scientific data using joint statistical moments

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2220587)