Matrix Profile XXIII: Contrast Profile: A Novel Time Series Primitive that Allows Real World Classification

Mercer, Ryan; Alaee, Sara; Abdoli, Alireza; Singh, Shailendra; Murillo, Amy C.; Keogh, Eamonn

doi:10.1109/icdm51629.2021.00151

Cited by 14 publications

(8 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For convinience, we refer to this length as attack length and denote it as L attack . Specifically, the adversary goal is to detect frequent patterns in time series, which is similar to motif discovery task [22,21,18,20] with one exception -the attacker could only use the shared Matrix Profile of length L util to detect the long motif.…”

Section: Threat Model and Attack Methodsmentioning

confidence: 99%

“…Two attack methods could retrieve some degree of long pattern information if MP is not protected. Note that existing motif discovery algorithms [22,21,18,20] could not be used to detect the long motif of length L attack through the shared Matrix Profile of length L util due to the lack of access to the original time series data. However, our proposed attack methods can still find long motif from a single shared MPI.…”

Section: Attack Methodsmentioning

confidence: 99%

See 1 more Smart Citation

PMP: Privacy-Aware Matrix Profile against Sensitive Pattern Inference for Time Series

Zhang¹,

Ding²,

Gao³

et al. 2023

Preprint

View full text Add to dashboard Cite

Recent rapid development of sensor technology has allowed massive fine-grained time series data to be collected and set foundation for the development of data-driven services and applications. During the process, data sharing is often involved to allow the third-party modelers to perform specific time series data mining tasks based on the need of data owner. The high resolution of time series data brings new challenges in protecting privacy. On one hand, meaningful information in high-resolution time series shifts from concrete point values to local shape-based segments. On the other hand, numerous research have found that long shape-based patterns could contain more sensitive information and may potentially be extracted and misused by a malicious third party. However, the privacy issue for time series patterns is surprisingly seldom explored in privacy-preserving literature. In this work, we consider a new privacy preserving problem: preventing malicious inference on long shape-based patterns while preserving short segment information for the utility task performance. To mitigate the challenge, we investigate an alternative approach by sharing Matrix Profile (MP), which is a nonlinear transformation of original data and a versatile data structure that supports many data mining tasks. We found that while MP can prevent the concrete shape leakage, the canonical correlation in MP index can still reveal the location of sensitive long pattern information. Based on this observation, we design two attacks named Location Attack and Entropy Attack to extract the pattern location from MP. To further protect MP from these two attacks, we propose a Privacy-Aware Matrix Profile (PMP) via perturbing the local correlation and breaking the canonical correlation in MP index vector. We evaluate our proposed PMP against baseline noise-adding methods through quantitative analysis and real-world case study to show the effectiveness of the proposed method. Our source code is available at https://github.com/lzhang18/PMP.

show abstract

Section: Threat Model and Attack Methodsmentioning

confidence: 99%

Section: Attack Methodsmentioning

confidence: 99%

PMP: Privacy-Aware Matrix Profile against Sensitive Pattern Inference for Time Series

Zhang¹,

Ding²,

Gao³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The Contrast Profile (Mercer et al 2021) is a recent tool for discovering contrasting patterns across time series, that is, behaviors that are repeated within one time series but are absent from another. Since the Contrast Profile is defined "lego-like", by combining several Matrix Profiles, this suggests that its definition could be retroactively generalized to Mplots.…”

Section: Exo25mentioning

confidence: 99%

Introducing Mplots: Scaling Time Series Recurrence Plots to Massive Datasets

Shahcheraghi,

Mercer,

Rodrigues

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Time series similarity matrices (informally, recurrence plots or dot-plots), are useful tools for time series data mining. They can be used to guide data exploration, and various useful features can be derived from them and then fed into downstream analytics. However, time series similarity matrices suffer from very poor scalability, taxing both time and memory requirements. In this work, we introduce novel ideas that allow us to scale the largest time series similarity matrices that can be examined by several orders of magnitude. The first idea is a novel algorithm to compute the matrices in a way that removes dependency on the subsequence length. This algorithm is so fast that it allows us to now address datasets where the memory limitations begin to dominate. Our second novel contribution is a multiscale algorithm that computes an approximation of the matrix appropriate for the limitations of the user’s memory/screen-resolution, then performs a local, just-in-time recomputation of any region that the user wishes to zoom-in on. Given that this largely removes time and space barriers, human visual attention then becomes the bottleneck. We further introduce algorithms that search massive matrices with quadrillions of cells and then prioritize regions for later examination by either humans or algorithms. We will demonstrate the utility of our ideas for data exploration, segmentation, and classification in domains as diverse as astronomy, bioinformatics, entomology, and wildlife monitoring.

show abstract

“…We will use the ubiquitous z-normalized Euclidean distance [34]. One minor modification to the Euclidean distance is that we clip it at √2𝑚 because values above this are anti-correlated in the Pearson Correlation space [20]. This is done in order to make the greatest use of the normalized range.…”

Section: Definitions and Notationmentioning

confidence: 99%

“…This definition resembles that of the recently introduced Contrast Profile [20], so we follow its normalization step which maximizes the effective normalization range by dividing 𝐄𝐏 𝑚 by √2𝑚, then clipping negative values to zero. This √2𝑚 factor makes the results unitless, and independent of both the sampling rate and subsequence length.…”

Section: Definitions and Notationmentioning

confidence: 99%

Novelets: A New Primitive that Allows Online Detection of Emerging Behaviors in Time Series

Mercer

Keogh

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Much of the world’s data is time series. While offline exploration of time series can be useful, time series is almost unique in allowing the possibility of direct and immediate intervention. For example, if we are monitoring an industrial process and have an algorithm that predicts imminent failure, we could direct a controller to open a pressure release valve or initiate an evacuation plan. There is a plethora of tools to monitor time series for known behaviors (pattern matching), previously unknown highly conserved behaviors (motifs), evolving behaviors (chains) and unexpected behaviors (anomalies). In this work, we claim that there is another useful primitive, emerging behaviors, that are worth monitoring for. We call such behaviors Novelets. We explain that Novelets are not anomalies, chains, or motifs but can be informally thought of as initially apparent anomalies that are later discovered to be motifs. We will show that Novelets have a natural interpretation in many disciplines, including science, medicine, and industry. As we will further demonstrate, Novelet discovery can have many downstream uses, including prognostics and abnormal behavior detection. We will demonstrate the utility of our proposed primitive on a diverse set of challenging domains.

show abstract

Matrix Profile XXIII: Contrast Profile: A Novel Time Series Primitive that Allows Real World Classification

Cited by 14 publications

References 8 publications

PMP: Privacy-Aware Matrix Profile against Sensitive Pattern Inference for Time Series

PMP: Privacy-Aware Matrix Profile against Sensitive Pattern Inference for Time Series

Introducing Mplots: Scaling Time Series Recurrence Plots to Massive Datasets

Novelets: A New Primitive that Allows Online Detection of Emerging Behaviors in Time Series

Contact Info

Product

Resources

About