Deep Reconstruction Error Based Unsupervised Outlier Detection in Time-Series. Amarbayasgalan, T., Lee, H. G., Van Huy, P., & Ryu, K. H. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), volume 12034 LNAI, pages 312–321, 2020. ISSN: 16113349doi abstract bibtex With all the advanced technology nowadays, the availability of time-series data is being increased. Outlier detection is an identification of abnormal patterns that provide useful information for many kinds of applications such as fraud detection, fault diagnosis, and disease detection. However, it will require an expensive domain and professional knowledge if there is no label which indicates normal and abnormality. Therefore, an unsupervised novelty detection approach will be used. In this paper, we propose a deep learning-based approach. First, it prepares subsequences according to the optimal lag length using Autoregressive (AR) model. The selected lag length for time-series analysis defines the data context in which further analysis is performed. Then, reconstruction errors (RE) of the subsequences on deep convolutional autoencoder (CAE) models are used to estimate the outlier threshold, and density-based clustering is used to identify outliers. We have compared the proposed method with several publicly available state-of-the-art anomaly detection methods on 30 time-series benchmark datasets. These results show that our proposed deep reconstruction error based approach outperforms the compared methods in most of the cases.
@inproceedings{Pham2020,
title = {Deep {Reconstruction} {Error} {Based} {Unsupervised} {Outlier} {Detection} in {Time}-{Series}},
volume = {12034 LNAI},
isbn = {978-3-030-42057-4},
doi = {10.1007/978-3-030-42058-1_26},
abstract = {With all the advanced technology nowadays, the availability of time-series data is being increased. Outlier detection is an identification of abnormal patterns that provide useful information for many kinds of applications such as fraud detection, fault diagnosis, and disease detection. However, it will require an expensive domain and professional knowledge if there is no label which indicates normal and abnormality. Therefore, an unsupervised novelty detection approach will be used. In this paper, we propose a deep learning-based approach. First, it prepares subsequences according to the optimal lag length using Autoregressive (AR) model. The selected lag length for time-series analysis defines the data context in which further analysis is performed. Then, reconstruction errors (RE) of the subsequences on deep convolutional autoencoder (CAE) models are used to estimate the outlier threshold, and density-based clustering is used to identify outliers. We have compared the proposed method with several publicly available state-of-the-art anomaly detection methods on 30 time-series benchmark datasets. These results show that our proposed deep reconstruction error based approach outperforms the compared methods in most of the cases.},
booktitle = {Lecture {Notes} in {Computer} {Science} (including subseries {Lecture} {Notes} in {Artificial} {Intelligence} and {Lecture} {Notes} in {Bioinformatics})},
author = {Amarbayasgalan, Tsatsral and Lee, Heon Gyu and Van Huy, Pham and Ryu, Keun Ho},
year = {2020},
note = {ISSN: 16113349},
keywords = {Anomaly, Autoregressive model, Deep convolutional autoencoder, Outlier, Time-series data},
pages = {312--321},
}
Downloads: 0
{"_id":"uMocsG2TTKPd7rnem","bibbaseid":"amarbayasgalan-lee-vanhuy-ryu-deepreconstructionerrorbasedunsupervisedoutlierdetectionintimeseries-2020","author_short":["Amarbayasgalan, T.","Lee, H. G.","Van Huy, P.","Ryu, K. H."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","title":"Deep Reconstruction Error Based Unsupervised Outlier Detection in Time-Series","volume":"12034 LNAI","isbn":"978-3-030-42057-4","doi":"10.1007/978-3-030-42058-1_26","abstract":"With all the advanced technology nowadays, the availability of time-series data is being increased. Outlier detection is an identification of abnormal patterns that provide useful information for many kinds of applications such as fraud detection, fault diagnosis, and disease detection. However, it will require an expensive domain and professional knowledge if there is no label which indicates normal and abnormality. Therefore, an unsupervised novelty detection approach will be used. In this paper, we propose a deep learning-based approach. First, it prepares subsequences according to the optimal lag length using Autoregressive (AR) model. The selected lag length for time-series analysis defines the data context in which further analysis is performed. Then, reconstruction errors (RE) of the subsequences on deep convolutional autoencoder (CAE) models are used to estimate the outlier threshold, and density-based clustering is used to identify outliers. We have compared the proposed method with several publicly available state-of-the-art anomaly detection methods on 30 time-series benchmark datasets. These results show that our proposed deep reconstruction error based approach outperforms the compared methods in most of the cases.","booktitle":"Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)","author":[{"propositions":[],"lastnames":["Amarbayasgalan"],"firstnames":["Tsatsral"],"suffixes":[]},{"propositions":[],"lastnames":["Lee"],"firstnames":["Heon","Gyu"],"suffixes":[]},{"propositions":[],"lastnames":["Van","Huy"],"firstnames":["Pham"],"suffixes":[]},{"propositions":[],"lastnames":["Ryu"],"firstnames":["Keun","Ho"],"suffixes":[]}],"year":"2020","note":"ISSN: 16113349","keywords":"Anomaly, Autoregressive model, Deep convolutional autoencoder, Outlier, Time-series data","pages":"312–321","bibtex":"@inproceedings{Pham2020,\n\ttitle = {Deep {Reconstruction} {Error} {Based} {Unsupervised} {Outlier} {Detection} in {Time}-{Series}},\n\tvolume = {12034 LNAI},\n\tisbn = {978-3-030-42057-4},\n\tdoi = {10.1007/978-3-030-42058-1_26},\n\tabstract = {With all the advanced technology nowadays, the availability of time-series data is being increased. Outlier detection is an identification of abnormal patterns that provide useful information for many kinds of applications such as fraud detection, fault diagnosis, and disease detection. However, it will require an expensive domain and professional knowledge if there is no label which indicates normal and abnormality. Therefore, an unsupervised novelty detection approach will be used. In this paper, we propose a deep learning-based approach. First, it prepares subsequences according to the optimal lag length using Autoregressive (AR) model. The selected lag length for time-series analysis defines the data context in which further analysis is performed. Then, reconstruction errors (RE) of the subsequences on deep convolutional autoencoder (CAE) models are used to estimate the outlier threshold, and density-based clustering is used to identify outliers. We have compared the proposed method with several publicly available state-of-the-art anomaly detection methods on 30 time-series benchmark datasets. These results show that our proposed deep reconstruction error based approach outperforms the compared methods in most of the cases.},\n\tbooktitle = {Lecture {Notes} in {Computer} {Science} (including subseries {Lecture} {Notes} in {Artificial} {Intelligence} and {Lecture} {Notes} in {Bioinformatics})},\n\tauthor = {Amarbayasgalan, Tsatsral and Lee, Heon Gyu and Van Huy, Pham and Ryu, Keun Ho},\n\tyear = {2020},\n\tnote = {ISSN: 16113349},\n\tkeywords = {Anomaly, Autoregressive model, Deep convolutional autoencoder, Outlier, Time-series data},\n\tpages = {312--321},\n}\n\n","author_short":["Amarbayasgalan, T.","Lee, H. G.","Van Huy, P.","Ryu, K. H."],"key":"Pham2020-1","id":"Pham2020-1","bibbaseid":"amarbayasgalan-lee-vanhuy-ryu-deepreconstructionerrorbasedunsupervisedoutlierdetectionintimeseries-2020","role":"author","urls":{},"keyword":["Anomaly","Autoregressive model","Deep convolutional autoencoder","Outlier","Time-series data"],"metadata":{"authorlinks":{}}},"bibtype":"inproceedings","biburl":"https://api.zotero.org/groups/2168152/items?key=VCdsaROd5deDY3prqqG8kI0c&format=bibtex&limit=100","dataSources":["syJjwTDDM32TsM2iF","QwrFbRJvXF69SEShv","HbngRCZLbLed2q9QT","LtEFvT85hYpNg4Esp","iHfnnAr7wKJJxkNMt","PrvBTxn4Zgeep29e5","78Yd9ZHcx783Wkffe","SKRhTA7ok4L4waPkZ","GfrMfnKTkYdcYTRsy","RqqCdXGEyWH4dZ76k","cbiwaQPQJSZeJDDY9","2Jak7xK39ytqcgqQ4","CDfDBPD6CDScj6Ty4","WgiCycoQjRx6KArBy","KBdipwowTNXWiKqYd","yjd6eECyb3TYZpZ3R","D9jmZ7aoHfJnYQ4ES","R8dLFAvyQ2oFRijDJ","dc6SzEK4S9LfC3XpA","kGWABmrDfhF29uibh","YE9GesxGLCsBc3vvC","v3qfuosZ66nvD85FK","BSxBG5ms26R2teZn9"],"keywords":["anomaly","autoregressive model","deep convolutional autoencoder","outlier","time-series data"],"search_terms":["deep","reconstruction","error","based","unsupervised","outlier","detection","time","series","amarbayasgalan","lee","van huy","ryu"],"title":"Deep Reconstruction Error Based Unsupervised Outlier Detection in Time-Series","year":2020}