R1SVM: a randomised nonlinear approach to large-scale anomaly detection

Erfani, Sarah M., Baktashmotlagh, Mahsa, Rajasegarar, Sutharshan, Karunasekera, Shanika and Leckie, Chris 2015, R1SVM: a randomised nonlinear approach to large-scale anomaly detection, in AAAI 2015: Proceedings of the Artificial Intelligence 2015 Conference, AAAI Publications, Palo Alto, Calif., pp. 432-438.

Attached Files
Name Description MIMEType Size Downloads

Title R1SVM: a randomised nonlinear approach to large-scale anomaly detection
Author(s) Erfani, Sarah M.
Baktashmotlagh, Mahsa
Rajasegarar, Sutharshan
Karunasekera, Shanika
Leckie, Chris
Conference name AAAI Conference on Artificial Intelligence: Workshop 15-14 (29th : 2015 : Austin, Texas)
Conference location Austin, Texas
Conference dates 25-30 Jan. 2015
Title of proceedings AAAI 2015: Proceedings of the Artificial Intelligence 2015 Conference
Editor(s) [Unknown]
Publication date 2015
Conference series AAAI Conference on Artificial Intelligence: Workshop 15-14
Start page 432
End page 438
Total pages 7
Publisher AAAI Publications
Place of publication Palo Alto, Calif.
Keyword(s) Anomaly detection
Outlier detection
One-class SVM
Random projection
Summary The problem of unsupervised anomaly detection arises in a wide variety of practical applications. While one-class support vector machines have demonstrated their effectiveness as an anomaly detection technique, their ability to model large datasets is limited due to their memory and time complexity for training. To address this issue for supervised learning of kernel machines, there has been growing interest in random projection methods as an alternative to the computationally expensive problems of kernel matrix construction and support vector optimisation. In this paper we leverage the theory of nonlinear random projections and propose the Randomised One-class SVM (R1SVM), which is an efficient and scalable anomaly detection technique that can be trained on large-scale datasets. Our empirical analysis on several real-life and synthetic datasets shows that our randomised 1SVM algorithm achieves comparable or better accuracy to deep autoen-coder and traditional kernelised approaches for anomaly detection, while being approximately 100 times faster in training and testing.
ISBN 9781577356998
ISSN 2374-3468
Language eng
Field of Research 080109 Pattern Recognition and Data Mining
Socio Economic Objective 890299 Computer Software and Services not elsewhere classified
HERDC Research category E1.1 Full written paper - refereed
ERA Research output type E Conference publication
Copyright notice ©2015, AAAI Publications
Persistent URL http://hdl.handle.net/10536/DRO/DU:30085019

Document type: Conference Paper
Collections: School of Information Technology
2018 ERA Submission
Connect to link resolver
Unless expressly stated otherwise, the copyright for items in DRO is owned by the author, with all rights reserved.

Version Filter Type
Citation counts: TR Web of Science Citation Count  Cited 0 times in TR Web of Science
Scopus Citation Count Cited 7 times in Scopus
Google Scholar Search Google Scholar
Access Statistics: 141 Abstract Views, 4 File Downloads  -  Detailed Statistics
Created: Thu, 21 Jul 2016, 10:49:41 EST

Every reasonable effort has been made to ensure that permission has been obtained for items included in DRO. If you believe that your rights have been infringed by this repository, please contact drosupport@deakin.edu.au.