dc.contributor.advisor | Kozat, Süleyman Serdar | |
dc.contributor.author | Gökcesu, Kaan | |
dc.date.accessioned | 2017-08-29T06:08:31Z | |
dc.date.available | 2017-08-29T06:08:31Z | |
dc.date.copyright | 2017-07 | |
dc.date.issued | 2017-08 | |
dc.date.submitted | 2017-08-28 | |
dc.identifier.uri | http://hdl.handle.net/11693/33561 | |
dc.description | Cataloged from PDF version of article. | en_US |
dc.description | Thesis (M.S.): Bilkent University, Department of Electrical and Electronics Engineering, İhsan Doğramacı Bilkent University, 2017 | en_US |
dc.description | Includes bibliographical references (leaves 64-71). | en_US |
dc.description.abstract | Online anomaly detection has attracted signi cant attention in recent years due
to its applications in network monitoring, cybersecurity, surveillance and sensor
failure. To this end, we introduce an algorithm that sequentially processes
data to detect anomalies in time series. Our algorithm consists of two stages:
density estimation and anomaly detection. First, we construct a probability density
function to model the normal data. Then, we threshold the density of the
newly observed data to detect anomalies. We approach this problem from an
information theoretic perspective and, for the rst time in the literature, propose
minimax optimal schemes for both stages to create an optimal anomaly detection
algorithm in a strong deterministic sense. For the rst stage, we introduce
an online density estimator that is minimax optimal for general nonstationary
exponential-family of distributions without any assumptions on the observation
sequence. Our algorithm does not require a priori knowledge of the time horizon,
the drift of the underlying distribution or the time instances the parameters of
the source changes. Our results are guaranteed to hold in an individual sequence
manner. For the second stage, we propose an online threshold selection scheme
that has logarithmic performance bounds against the best threshold chosen in
hindsight. Our complete algorithm adaptively updates its parameters in a truly
sequential manner to achieve log-linear regrets in both stages. Because of its
universal prediction perspective on its density estimation, our anomaly detection
algorithm can be used in unsupervised, semi-supervised and supervised manner.
Through synthetic and real life experiments, we demonstrate substantial performance
gains with respect to the state-of-the-art. | en_US |
dc.description.statementofresponsibility | by Kaan Gökcesu. | en_US |
dc.format.extent | ix, 71 leaves : charts (some color) ; 29 cm. | en_US |
dc.language.iso | English | en_US |
dc.rights | info:eu-repo/semantics/openAccess | en_US |
dc.subject | Anomaly Detection | en_US |
dc.subject | Time Series | en_US |
dc.subject | Online Learning | en_US |
dc.subject | Density Estimation | en_US |
dc.subject | Minimax Optimal | en_US |
dc.subject | Nonstationary | en_US |
dc.title | Online minimax optimal density estimation and anomaly detection in nonstationary environments | en_US |
dc.title.alternative | Durağan olmayan ortamlarda çevrimiçi minimaks optimal yoğunluk tahmini ve anomali tespiti | en_US |
dc.type | Thesis | en_US |
dc.department | Department of Electrical and Electronics Engineering | en_US |
dc.publisher | Bilkent University | en_US |
dc.description.degree | M.S. | en_US |
dc.identifier.itemid | B156121 | |
dc.embargo.release | 2020-08-25 | |