Change Detection in Multivariate data streams: Online Analysis with Kernel-QuantTree

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Notarianni, Michelangelo Olmo Nogara, Leveni, Filippo, Stucchi, Diego, Frittoli, Luca, Boracchi, Giacomo
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908380923166720
author Notarianni, Michelangelo Olmo Nogara
Leveni, Filippo
Stucchi, Diego
Frittoli, Luca
Boracchi, Giacomo
author_facet Notarianni, Michelangelo Olmo Nogara
Leveni, Filippo
Stucchi, Diego
Frittoli, Luca
Boracchi, Giacomo
contents We present Kernel-QuantTree Exponentially Weighted Moving Average (KQT-EWMA), a non-parametric change-detection algorithm that combines the Kernel-QuantTree (KQT) histogram and the EWMA statistic to monitor multivariate data streams online. The resulting monitoring scheme is very flexible, since histograms can be used to model any stationary distribution, and practical, since the distribution of test statistics does not depend on the distribution of datastream in stationary conditions (non-parametric monitoring). KQT-EWMA enables controlling false alarms by operating at a pre-determined Average Run Length ($ARL_0$), which measures the expected number of stationary samples to be monitored before triggering a false alarm. The latter peculiarity is in contrast with most non-parametric change-detection tests, which rarely can control the $ARL_0$ a priori. Our experiments on synthetic and real-world datasets demonstrate that KQT-EWMA can control $ARL_0$ while achieving detection delays comparable to or lower than state-of-the-art methods designed to work in the same conditions.
format Preprint
id arxiv_https___arxiv_org_abs_2410_13778
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Change Detection in Multivariate data streams: Online Analysis with Kernel-QuantTree
Notarianni, Michelangelo Olmo Nogara
Leveni, Filippo
Stucchi, Diego
Frittoli, Luca
Boracchi, Giacomo
Machine Learning
Artificial Intelligence
We present Kernel-QuantTree Exponentially Weighted Moving Average (KQT-EWMA), a non-parametric change-detection algorithm that combines the Kernel-QuantTree (KQT) histogram and the EWMA statistic to monitor multivariate data streams online. The resulting monitoring scheme is very flexible, since histograms can be used to model any stationary distribution, and practical, since the distribution of test statistics does not depend on the distribution of datastream in stationary conditions (non-parametric monitoring). KQT-EWMA enables controlling false alarms by operating at a pre-determined Average Run Length ($ARL_0$), which measures the expected number of stationary samples to be monitored before triggering a false alarm. The latter peculiarity is in contrast with most non-parametric change-detection tests, which rarely can control the $ARL_0$ a priori. Our experiments on synthetic and real-world datasets demonstrate that KQT-EWMA can control $ARL_0$ while achieving detection delays comparable to or lower than state-of-the-art methods designed to work in the same conditions.
title Change Detection in Multivariate data streams: Online Analysis with Kernel-QuantTree
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2410.13778