Unsupervised real-time anomaly detection for streaming data. Ahmad, S., Lavin, A., Purdy, S., & Agha, Z. Neurocomputing, 262:134–147, November, 2017. ![Unsupervised real-time anomaly detection for streaming data [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex We are seeing an enormous increase in the availability of streaming, time-series data. Largely driven by the rise of connected real-time data sources, this data presents technical challenges and opportunities. One fundamental capability for streaming analytics is to model each stream in an unsupervised fashion and detect unusual, anomalous behaviors in real-time. Early anomaly detection is valuable, yet it can be difficult to execute reliably in practice. Application constraints require systems to process data in real-time, not batches. Streaming data inherently exhibits concept drift, favoring algorithms that learn continuously. Furthermore, the massive number of independent streams in practice requires that anomaly detectors be fully automated. In this paper we propose a novel anomaly detection algorithm that meets these constraints. The technique is based on an online sequence memory algorithm called Hierarchical Temporal Memory (HTM). We also present results using the Numenta Anomaly Benchmark (NAB), a benchmark containing real-world data streams with labeled anomalies. The benchmark, the first of its kind, provides a controlled open-source environment for testing anomaly detection algorithms on streaming data. We present results and analysis for a wide range of algorithms on this benchmark, and discuss future challenges for the emerging field of streaming analytics.
@article{ahmad_unsupervised_2017,
series = {Online {Real}-{Time} {Learning} {Strategies} for {Data} {Streams}},
title = {Unsupervised real-time anomaly detection for streaming data},
volume = {262},
issn = {0925-2312},
url = {http://www.sciencedirect.com/science/article/pii/S0925231217309864},
doi = {10.1016/j.neucom.2017.04.070},
abstract = {We are seeing an enormous increase in the availability of streaming, time-series data. Largely driven by the rise of connected real-time data sources, this data presents technical challenges and opportunities. One fundamental capability for streaming analytics is to model each stream in an unsupervised fashion and detect unusual, anomalous behaviors in real-time. Early anomaly detection is valuable, yet it can be difficult to execute reliably in practice. Application constraints require systems to process data in real-time, not batches. Streaming data inherently exhibits concept drift, favoring algorithms that learn continuously. Furthermore, the massive number of independent streams in practice requires that anomaly detectors be fully automated. In this paper we propose a novel anomaly detection algorithm that meets these constraints. The technique is based on an online sequence memory algorithm called Hierarchical Temporal Memory (HTM). We also present results using the Numenta Anomaly Benchmark (NAB), a benchmark containing real-world data streams with labeled anomalies. The benchmark, the first of its kind, provides a controlled open-source environment for testing anomaly detection algorithms on streaming data. We present results and analysis for a wide range of algorithms on this benchmark, and discuss future challenges for the emerging field of streaming analytics.},
language = {en},
urldate = {2020-12-21},
journal = {Neurocomputing},
author = {Ahmad, Subutai and Lavin, Alexander and Purdy, Scott and Agha, Zuha},
month = nov,
year = {2017},
keywords = {Anomaly detection, Benchmark dataset, Concept drift, Hierarchical Temporal Memory, Streaming data, Unsupervised learning},
pages = {134--147},
}
Downloads: 0
{"_id":"ZAcw35aJiZGtZsaXZ","bibbaseid":"ahmad-lavin-purdy-agha-unsupervisedrealtimeanomalydetectionforstreamingdata-2017","author_short":["Ahmad, S.","Lavin, A.","Purdy, S.","Agha, Z."],"bibdata":{"bibtype":"article","type":"article","series":"Online Real-Time Learning Strategies for Data Streams","title":"Unsupervised real-time anomaly detection for streaming data","volume":"262","issn":"0925-2312","url":"http://www.sciencedirect.com/science/article/pii/S0925231217309864","doi":"10.1016/j.neucom.2017.04.070","abstract":"We are seeing an enormous increase in the availability of streaming, time-series data. Largely driven by the rise of connected real-time data sources, this data presents technical challenges and opportunities. One fundamental capability for streaming analytics is to model each stream in an unsupervised fashion and detect unusual, anomalous behaviors in real-time. Early anomaly detection is valuable, yet it can be difficult to execute reliably in practice. Application constraints require systems to process data in real-time, not batches. Streaming data inherently exhibits concept drift, favoring algorithms that learn continuously. Furthermore, the massive number of independent streams in practice requires that anomaly detectors be fully automated. In this paper we propose a novel anomaly detection algorithm that meets these constraints. The technique is based on an online sequence memory algorithm called Hierarchical Temporal Memory (HTM). We also present results using the Numenta Anomaly Benchmark (NAB), a benchmark containing real-world data streams with labeled anomalies. The benchmark, the first of its kind, provides a controlled open-source environment for testing anomaly detection algorithms on streaming data. We present results and analysis for a wide range of algorithms on this benchmark, and discuss future challenges for the emerging field of streaming analytics.","language":"en","urldate":"2020-12-21","journal":"Neurocomputing","author":[{"propositions":[],"lastnames":["Ahmad"],"firstnames":["Subutai"],"suffixes":[]},{"propositions":[],"lastnames":["Lavin"],"firstnames":["Alexander"],"suffixes":[]},{"propositions":[],"lastnames":["Purdy"],"firstnames":["Scott"],"suffixes":[]},{"propositions":[],"lastnames":["Agha"],"firstnames":["Zuha"],"suffixes":[]}],"month":"November","year":"2017","keywords":"Anomaly detection, Benchmark dataset, Concept drift, Hierarchical Temporal Memory, Streaming data, Unsupervised learning","pages":"134–147","bibtex":"@article{ahmad_unsupervised_2017,\n\tseries = {Online {Real}-{Time} {Learning} {Strategies} for {Data} {Streams}},\n\ttitle = {Unsupervised real-time anomaly detection for streaming data},\n\tvolume = {262},\n\tissn = {0925-2312},\n\turl = {http://www.sciencedirect.com/science/article/pii/S0925231217309864},\n\tdoi = {10.1016/j.neucom.2017.04.070},\n\tabstract = {We are seeing an enormous increase in the availability of streaming, time-series data. Largely driven by the rise of connected real-time data sources, this data presents technical challenges and opportunities. One fundamental capability for streaming analytics is to model each stream in an unsupervised fashion and detect unusual, anomalous behaviors in real-time. Early anomaly detection is valuable, yet it can be difficult to execute reliably in practice. Application constraints require systems to process data in real-time, not batches. Streaming data inherently exhibits concept drift, favoring algorithms that learn continuously. Furthermore, the massive number of independent streams in practice requires that anomaly detectors be fully automated. In this paper we propose a novel anomaly detection algorithm that meets these constraints. The technique is based on an online sequence memory algorithm called Hierarchical Temporal Memory (HTM). We also present results using the Numenta Anomaly Benchmark (NAB), a benchmark containing real-world data streams with labeled anomalies. The benchmark, the first of its kind, provides a controlled open-source environment for testing anomaly detection algorithms on streaming data. We present results and analysis for a wide range of algorithms on this benchmark, and discuss future challenges for the emerging field of streaming analytics.},\n\tlanguage = {en},\n\turldate = {2020-12-21},\n\tjournal = {Neurocomputing},\n\tauthor = {Ahmad, Subutai and Lavin, Alexander and Purdy, Scott and Agha, Zuha},\n\tmonth = nov,\n\tyear = {2017},\n\tkeywords = {Anomaly detection, Benchmark dataset, Concept drift, Hierarchical Temporal Memory, Streaming data, Unsupervised learning},\n\tpages = {134--147},\n}\n\n\n\n","author_short":["Ahmad, S.","Lavin, A.","Purdy, S.","Agha, Z."],"key":"ahmad_unsupervised_2017","id":"ahmad_unsupervised_2017","bibbaseid":"ahmad-lavin-purdy-agha-unsupervisedrealtimeanomalydetectionforstreamingdata-2017","role":"author","urls":{"Paper":"http://www.sciencedirect.com/science/article/pii/S0925231217309864"},"keyword":["Anomaly detection","Benchmark dataset","Concept drift","Hierarchical Temporal Memory","Streaming data","Unsupervised learning"],"metadata":{"authorlinks":{}},"html":""},"bibtype":"article","biburl":"https://bibbase.org/zotero/mh_lenguyen","dataSources":["iwKepCrWBps7ojhDx"],"keywords":["anomaly detection","benchmark dataset","concept drift","hierarchical temporal memory","streaming data","unsupervised learning"],"search_terms":["unsupervised","real","time","anomaly","detection","streaming","data","ahmad","lavin","purdy","agha"],"title":"Unsupervised real-time anomaly detection for streaming data","year":2017}