A Benchmarking Study of Unsupervised Anomaly Detection Algorithms

Graduate Thesis Or Dissertation

Public Deposited

Download PDF

Citeable URL: https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/sn00b4999

Descriptions

Attribute Name	Values
Creator	Emmott, Andrew F.
Abstract	It is common practice in the unsupervised anomaly detection literature to create experimental benchmarks by sampling from existing supervised learning datasets. We seek to improve this practice by identifying four dimensions important to real-world anomaly detection applications --- point difficulty, clusteredness of anomalies, relevance of features, and relative frequency of anomalies --- and then proposing how to simulate and control these factors when sampling points during benchmark creation. We apply this methodology to produce a large corpus of unsupervised anomaly detection benchmarks and then evaluate several state-of-the-art anomaly detection algorithms against this corpus. Our final analysis not only compares the performance of these algorithms across a large variety of problems, but it also assesses the impact of our identified problems dimensions on experimental outcomes.
License	All rights reserved
Resource Type	Masters Thesis
Date Issued	2020-06-18
Degree Level	Master's
Degree Name	Master of Science (M.S.)
Degree Field	Computer Science
Degree Grantor	Oregon State University
Commencement Year	2021
Advisor	Dietterich, Thomas G.
Committee Member	Borradaile, Glencora Tadepalli, Prasad
Academic Affiliation	Electrical Engineering and Computer Science
Subject	Benchmarking (Management) Anomaly detection (Computer security)
Rights Statement	In Copyright
Publisher	Oregon State University
Peer Reviewed	No
Language	English [eng]

Parents:

This work has no parents.

In Collection:

Thumbnail	Title	Date Uploaded	Visibility	Actions
	EmmottAndrew2020.pdf	2020-07-02	Public	Download