Is It Worth It? Comparing Six Deep and Classical Methods for Unsupervised Anomaly Detection in Time Series. Rewicki, F., Denzler, J., & Niebling, J. Applied Sciences, 13(3):1778, January, 2023. Number: 3 Publisher: Multidisciplinary Digital Publishing Institute![Is It Worth It? Comparing Six Deep and Classical Methods for Unsupervised Anomaly Detection in Time Series [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex Detecting anomalies in time series data is important in a variety of fields, including system monitoring, healthcare and cybersecurity. While the abundance of available methods makes it difficult to choose the most appropriate method for a given application, each method has its strengths in detecting certain types of anomalies. In this study, we compare six unsupervised anomaly detection methods of varying complexity to determine whether more complex methods generally perform better and if certain methods are better suited to certain types of anomalies. We evaluated the methods using the UCR anomaly archive, a recent benchmark dataset for anomaly detection. We analyzed the results on a dataset and anomaly-type level after adjusting the necessary hyperparameters for each method. Additionally, we assessed the ability of each method to incorporate prior knowledge about anomalies and examined the differences between point-wise and sequence-wise features. Our experiments show that classical machine learning methods generally outperform deep learning methods across a range of anomaly types.
@article{rewicki_is_2023,
title = {Is {It} {Worth} {It}? {Comparing} {Six} {Deep} and {Classical} {Methods} for {Unsupervised} {Anomaly} {Detection} in {Time} {Series}},
volume = {13},
copyright = {http://creativecommons.org/licenses/by/3.0/},
issn = {2076-3417},
shorttitle = {Is {It} {Worth} {It}?},
url = {https://www.mdpi.com/2076-3417/13/3/1778},
doi = {10.3390/app13031778},
abstract = {Detecting anomalies in time series data is important in a variety of fields, including system monitoring, healthcare and cybersecurity. While the abundance of available methods makes it difficult to choose the most appropriate method for a given application, each method has its strengths in detecting certain types of anomalies. In this study, we compare six unsupervised anomaly detection methods of varying complexity to determine whether more complex methods generally perform better and if certain methods are better suited to certain types of anomalies. We evaluated the methods using the UCR anomaly archive, a recent benchmark dataset for anomaly detection. We analyzed the results on a dataset and anomaly-type level after adjusting the necessary hyperparameters for each method. Additionally, we assessed the ability of each method to incorporate prior knowledge about anomalies and examined the differences between point-wise and sequence-wise features. Our experiments show that classical machine learning methods generally outperform deep learning methods across a range of anomaly types.},
language = {en},
number = {3},
urldate = {2023-11-14},
journal = {Applied Sciences},
author = {Rewicki, Ferdinand and Denzler, Joachim and Niebling, Julia},
month = jan,
year = {2023},
note = {Number: 3
Publisher: Multidisciplinary Digital Publishing Institute},
keywords = {Anomaly detection, Time series, Deep Learning, Benchmark, Machine learning},
pages = {1778},
file = {Full Text PDF:C\:\\Users\\Guillaume\\Zotero\\storage\\IU3IXMZ2\\Rewicki et al. - 2023 - Is It Worth It Comparing Six Deep and Classical M.pdf:application/pdf},
}
Downloads: 0
{"_id":"2M3sxdwGDHuwv7R5q","bibbaseid":"rewicki-denzler-niebling-isitworthitcomparingsixdeepandclassicalmethodsforunsupervisedanomalydetectionintimeseries-2023","author_short":["Rewicki, F.","Denzler, J.","Niebling, J."],"bibdata":{"bibtype":"article","type":"article","title":"Is It Worth It? Comparing Six Deep and Classical Methods for Unsupervised Anomaly Detection in Time Series","volume":"13","copyright":"http://creativecommons.org/licenses/by/3.0/","issn":"2076-3417","shorttitle":"Is It Worth It?","url":"https://www.mdpi.com/2076-3417/13/3/1778","doi":"10.3390/app13031778","abstract":"Detecting anomalies in time series data is important in a variety of fields, including system monitoring, healthcare and cybersecurity. While the abundance of available methods makes it difficult to choose the most appropriate method for a given application, each method has its strengths in detecting certain types of anomalies. In this study, we compare six unsupervised anomaly detection methods of varying complexity to determine whether more complex methods generally perform better and if certain methods are better suited to certain types of anomalies. We evaluated the methods using the UCR anomaly archive, a recent benchmark dataset for anomaly detection. We analyzed the results on a dataset and anomaly-type level after adjusting the necessary hyperparameters for each method. Additionally, we assessed the ability of each method to incorporate prior knowledge about anomalies and examined the differences between point-wise and sequence-wise features. Our experiments show that classical machine learning methods generally outperform deep learning methods across a range of anomaly types.","language":"en","number":"3","urldate":"2023-11-14","journal":"Applied Sciences","author":[{"propositions":[],"lastnames":["Rewicki"],"firstnames":["Ferdinand"],"suffixes":[]},{"propositions":[],"lastnames":["Denzler"],"firstnames":["Joachim"],"suffixes":[]},{"propositions":[],"lastnames":["Niebling"],"firstnames":["Julia"],"suffixes":[]}],"month":"January","year":"2023","note":"Number: 3 Publisher: Multidisciplinary Digital Publishing Institute","keywords":"Anomaly detection, Time series, Deep Learning, Benchmark, Machine learning","pages":"1778","file":"Full Text PDF:C\\:\\\\Users\\\\Guillaume\\\\Zotero\\\\storage\\\\IU3IXMZ2\\\\Rewicki et al. - 2023 - Is It Worth It Comparing Six Deep and Classical M.pdf:application/pdf","bibtex":"@article{rewicki_is_2023,\n\ttitle = {Is {It} {Worth} {It}? {Comparing} {Six} {Deep} and {Classical} {Methods} for {Unsupervised} {Anomaly} {Detection} in {Time} {Series}},\n\tvolume = {13},\n\tcopyright = {http://creativecommons.org/licenses/by/3.0/},\n\tissn = {2076-3417},\n\tshorttitle = {Is {It} {Worth} {It}?},\n\turl = {https://www.mdpi.com/2076-3417/13/3/1778},\n\tdoi = {10.3390/app13031778},\n\tabstract = {Detecting anomalies in time series data is important in a variety of fields, including system monitoring, healthcare and cybersecurity. While the abundance of available methods makes it difficult to choose the most appropriate method for a given application, each method has its strengths in detecting certain types of anomalies. In this study, we compare six unsupervised anomaly detection methods of varying complexity to determine whether more complex methods generally perform better and if certain methods are better suited to certain types of anomalies. We evaluated the methods using the UCR anomaly archive, a recent benchmark dataset for anomaly detection. We analyzed the results on a dataset and anomaly-type level after adjusting the necessary hyperparameters for each method. Additionally, we assessed the ability of each method to incorporate prior knowledge about anomalies and examined the differences between point-wise and sequence-wise features. Our experiments show that classical machine learning methods generally outperform deep learning methods across a range of anomaly types.},\n\tlanguage = {en},\n\tnumber = {3},\n\turldate = {2023-11-14},\n\tjournal = {Applied Sciences},\n\tauthor = {Rewicki, Ferdinand and Denzler, Joachim and Niebling, Julia},\n\tmonth = jan,\n\tyear = {2023},\n\tnote = {Number: 3\nPublisher: Multidisciplinary Digital Publishing Institute},\n\tkeywords = {Anomaly detection, Time series, Deep Learning, Benchmark, Machine learning},\n\tpages = {1778},\n\tfile = {Full Text PDF:C\\:\\\\Users\\\\Guillaume\\\\Zotero\\\\storage\\\\IU3IXMZ2\\\\Rewicki et al. - 2023 - Is It Worth It Comparing Six Deep and Classical M.pdf:application/pdf},\n}\n\n","author_short":["Rewicki, F.","Denzler, J.","Niebling, J."],"key":"rewicki_is_2023","id":"rewicki_is_2023","bibbaseid":"rewicki-denzler-niebling-isitworthitcomparingsixdeepandclassicalmethodsforunsupervisedanomalydetectionintimeseries-2023","role":"author","urls":{"Paper":"https://www.mdpi.com/2076-3417/13/3/1778"},"keyword":["Anomaly detection","Time series","Deep Learning","Benchmark","Machine learning"],"metadata":{"authorlinks":{}},"html":""},"bibtype":"article","biburl":"https://bibbase.org/network/files/taiAEmmZxo42FqfQN","dataSources":["qG4exu4XyJLWyPN7m"],"keywords":["anomaly detection","time series","deep learning","benchmark","machine learning"],"search_terms":["worth","comparing","six","deep","classical","methods","unsupervised","anomaly","detection","time","series","rewicki","denzler","niebling"],"title":"Is It Worth It? Comparing Six Deep and Classical Methods for Unsupervised Anomaly Detection in Time Series","year":2023}