Real time change point detection by incremental PCA in large scale sensor data

Mishin, Dmitry; Brantner-Magee, Kieran; Czako, Ferenc; Szalay, Alexander S.

doi:10.1109/hpec.2014.7040959

Cited by 4 publications

(9 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another common approach to dimensionality reduction is principal component analysis (PCA) [28], which achieves dimensionality reduction by projecting the signal along the singular space of the leading singular values. In this case, A or A t corresponds to the signal singular space.…”

Section: Related Workmentioning

confidence: 99%

“…(iv) (PCA). There are also approaches to change-point detection using principal component analysis (PCA) of the data streams (e.g., [28,34]), which can be viewed as using a deterministic fixed projection A, which is pre-computed as the signal singular space associated with the leading singular values of the data covariance matrix. (v) (Missing data).…”

Section: Sketching Matricesmentioning

confidence: 99%

See 1 more Smart Citation

Sketching for sequential change-point detection

Cao

Thompson

Wang

et al. 2019

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

We present sequential change-point detection procedures based on linear sketches of high-dimensional signal vectors using generalized likelihood ratio (GLR) statistics. The GLR statistics allow for an unknown post-change mean that represents an anomaly or novelty. We consider both fixed and time-varying projections, derive theoretical approximations to two fundamental performance metrics: the average run length (ARL) and the expected detection delay (EDD); these approximations are shown to be highly accurate by numerical simulations. We further characterize the relative performance measure of the sketching procedure compared to that without sketching and show that there can be little performance loss when the signal strength is sufficiently large, and enough number of sketches are used. Finally, we demonstrate the good performance of sketching procedures using simulation and real-data examples on solar flare detection and failure detection in power networks.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Sketching Matricesmentioning

confidence: 99%

Sketching for sequential change-point detection

Cao

Thompson

Wang

et al. 2019

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

show abstract

“…And, most importantly, we do not use the entire residual subspace, but rather the subspace of it that is most sensitive to a user-defined set of relevant distributional changes. Other examples of PCA-based anomaly detection procedures are Ferrer (2007), Mishin et al (2014) and Harrou et al (2015). None of the articles mentioned in this paragraph considers the speed of detection, which is a major difference to our objective.…”

Section: Connections With Prior Workmentioning

confidence: 99%

“…In many applications, these sensors are deployed in large networks for online monitoring of a system. Concrete examples include temperature monitoring of a data center at Johns Hopkins (Mishin et al 2014), plant-wide monitoring of industrial processes (Ge 2017) and semiconductor manufacturing (Zou et al 2014). Similar technology is also used within video segmentation (Kuncheva and Faithfull 2014), solar flare detection (Liu et al 2015), medical monitoring, DNA protein sequence analysis, network intrusion detection and speech recognition.…”

Section: Introduction 11 Motivationmentioning

confidence: 99%

Online Detection of Sparse Changes in High-Dimensional Data Streams Using Tailored Projections

Tveten¹,

Glad²

2019

Preprint

View full text Add to dashboard Cite

When applying principal component analysis (PCA) for dimension reduction, the most varying projections are usually used in order to retain most of the information. For the purpose of anomaly and change detection, however, the least varying projections are often the most important ones. In this article, we present a novel method that automatically tailors the choice of projections to monitor for sparse changes in the mean and/or covariance matrix of high-dimensional data. A subset of the least varying projections is almost always selected based on a criteria of the projection's sensitivity to changes.Our focus is on online/sequential change detection, where the aim is to detect changes as quickly as possible, while controlling false alarms at a specified level. A combination of tailored PCA and a generalized log-likelihood monitoring procedure displays high efficiency in detecting even very sparse changes in the mean, variance and correlation. We demonstrate on real data that tailored PCA monitoring is efficient for sparse change detection also when the data streams are highly auto-correlated and non-normal. Notably, error control is achieved without a large validation set, which is needed in most existing methods.

show abstract

“…Using PCA in SPC goes back to the work of Jackson and Morris () and Jackson and Mudholkar (), and its various extensions (see Ketelaere et al, and Rato et al, , for an overview) have been successfully applied to many real data situations. Within the machine learning literature on anomaly detection, Mishin et al () use PCA for temperature monitoring at Johns Hopkins, Harrou et al () apply PCA‐based anomaly detection to find segments with abnormal rates of patient arrivals at an emergency department, and Camacho et al () relate PCA‐based monitoring in SPC to modern anomaly detection in statistical networks. PCA has also been studied in the setting of change detection in multivariate functional data with the aim of detecting faulty profiles in a forging manufacturing process (Wang et al, ).…”

Section: Introductionmentioning

confidence: 99%

Which principal components are most sensitive in the change detection problem?

Tveten

2019

Stat

View full text Add to dashboard Cite

Principal component analysis (PCA) is often used in anomaly detection and statistical process control tasks. For bivariate normal data, we prove that the minor projection (the least varying projection) of the PCA‐rotated data is the most sensitive to distributional changes, where sensitivity is defined as the Hellinger distance between the projections' marginal distributions before and after a change. In particular, this is almost always the case if only one parameter of the bivariate normal distribution changes, that is, the change is sparse. Simulations indicate that the minor projections are the most sensitive for a large range of changes and pre‐change settings in higher dimensions as well, including changes that are very sparse. This motivates using only a few of the minor projections for detecting sparse distributional changes in high‐dimensional data.

show abstract

Real time change point detection by incremental PCA in large scale sensor data

Cited by 4 publications

References 4 publications

Sketching for sequential change-point detection

Sketching for sequential change-point detection

Online Detection of Sparse Changes in High-Dimensional Data Streams Using Tailored Projections

Which principal components are most sensitive in the change detection problem?

Contact Info

Product

Resources

About