An outlier in my ocean hydrology model.
Working out when something you are seeing is not something you were expecting, when you are not quite sure how to quantify what you expect actually. Detecting black swans, for example.
Nuit Blanche rounds up methods for high dimensional statistics.
Trend detection
e.g. Gnip Trend detection Nikolov (2012) looks fun, and comes with a cute explanation of how you might do this nonparametrically.
Incoming
Sevvandi Kandanaarachchi is presenting some methods in a seminar right now: (Kandanaarachchi and Hyndman 2021; Kandanaarachchi, Hyndman, and Smith-Miles 2020; Kandanaarachchi 2021; Sevvandi Kandanaarachchi and Hyndman 2021; Talagala et al. 2020).
Related software:
References
Ahmad, Subutai, Alexander Lavin, Scott Purdy, and Zuha Agha. 2017. βUnsupervised Real-Time Anomaly Detection for Streaming Data.β Neurocomputing, Online Real-Time Learning Strategies for Data Streams, 262 (November): 134β47.
Chalapathy, Raghavendra, and Sanjay Chawla. 2019. βDeep Learning for Anomaly Detection: A Survey.β arXiv:1901.03407 [Cs, Stat], January.
Henrickson, Scott, Jeff Kolb, Brian Lehmann, and Joshua Montague. 2015. βTrend Detection in Social Data.β Twitter.
Kandanaarachchi, Sevvandi. 2021. βUnsupervised Anomaly Detection Ensembles Using Item Response Theory.β arXiv:2106.06243 [Cs, Stat], June.
Kandanaarachchi, Sevvandi, and Rob J. Hyndman. 2021. βDimension Reduction for Outlier Detection Using DOBIN.β Journal of Computational and Graphical Statistics 30 (1): 204β19.
Kandanaarachchi, Sevvandi, Rob J. Hyndman, and Kate Smith-Miles. 2020. βEarly Classification of Spatio-Temporal Events Using Partial Information.β PLOS ONE 15 (8): e0236331.
Kandanaarachchi, Sevvandi, Mario A. Munoz, Rob J. Hyndman, and Kate Smith-Miles. 2018. βOn Normalization and Algorithm Selection for Unsupervised Outlier Detection.β 16/18. Monash Econometrics and Business Statistics Working Papers. Monash University, Department of Econometrics and Business Statistics.
Kristiadi, Agustinus, Matthias Hein, and Philipp Hennig. 2022. βBeing a Bit Frequentist Improves Bayesian Neural Networks.β In CoRR. arXiv.
Kulinski, Sean, and David I. Inouye. 2022. βTowards Explaining Distribution Shifts.β arXiv.
Lopez, Jose A., Octavia Camps, and Mario Sznaier. 2015. βRobust Anomaly Detection Using Semidefinite Programming.β arXiv:1504.00905 [Cs, Math], April.
Ming, Yifei, Ying Fan, and Yixuan Li. 2022. βPOEM: Out-of-Distribution Detection with Posterior Sampling,β June.
Nikolov, Stanislav. 2012. βTrend or no trend : a novel nonparametric method for classifying time series.β Thesis, Massachusetts Institute of Technology.
Ovadia, Yaniv, Emily Fertig, Jie Ren, Zachary Nado, D. Sculley, Sebastian Nowozin, Joshua V. Dillon, Balaji Lakshminarayanan, and Jasper Snoek. 2019. βCan You Trust Your Modelβs Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shift.β In Proceedings of the 33rd International Conference on Neural Information Processing Systems, 14003β14. Red Hook, NY, USA: Curran Associates Inc.
Pang, Guansong, Chunhua Shen, Longbing Cao, and Anton van den Hengel. 2021. βDeep Learning for Anomaly Detection: A Review.β ACM Computing Surveys 54 (2): 1β38.
Sevvandi Kandanaarachchi, and Rob J Hyndman. 2021. βLeave-One-Out Kernel Density Estimates for Outlier Detection.β
She, Yiyuan, and Art B. Owen. 2010. βOutlier Detection Using Nonconvex Penalized Regression.β
Talagala, Priyanga Dilini, Rob J. Hyndman, Kate Smith-Miles, Sevvandi Kandanaarachchi, and Mario A. MuΓ±oz. 2020. βAnomaly Detection in Streaming Nonstationary Temporal Data.β Journal of Computational and Graphical Statistics 29 (1): 13β27.
No comments yet. Why not leave one?