Isolation Forest
Most existing model-based approaches to anomaly de-tection construct a profile of normal instances, then iden-tify instances that do not conform to the normal profile as anomalies. This paper proposes a fundamentally different model-based method that explicitly isolates anomalies in-stead of profile...
Main Authors: | , , |
---|---|
Other Authors: | |
Format: | Text |
Language: | English |
Subjects: | |
Online Access: | http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.678.3903 http://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/icdm08b.pdf |
id |
ftciteseerx:oai:CiteSeerX.psu:10.1.1.678.3903 |
---|---|
record_format |
openpolar |
spelling |
ftciteseerx:oai:CiteSeerX.psu:10.1.1.678.3903 2023-05-15T17:53:46+02:00 Isolation Forest Fei Tony Liu Kai Ming Ting Zhi-hua Zhou The Pennsylvania State University CiteSeerX Archives application/pdf http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.678.3903 http://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/icdm08b.pdf en eng http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.678.3903 http://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/icdm08b.pdf Metadata may be used without restrictions as long as the oai identifier remains attached to it. http://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/icdm08b.pdf text ftciteseerx 2016-01-08T17:43:42Z Most existing model-based approaches to anomaly de-tection construct a profile of normal instances, then iden-tify instances that do not conform to the normal profile as anomalies. This paper proposes a fundamentally different model-based method that explicitly isolates anomalies in-stead of profiles normal points. To our best knowledge, the concept of isolation has not been explored in current liter-ature. The use of isolation enables the proposed method, iForest, to exploit sub-sampling to an extent that is not fea-sible in existing methods, creating an algorithm which has a linear time complexity with a low constant and a low mem-ory requirement. Our empirical evaluation shows that iFor-est performs favourably to ORCA, a near-linear time com-plexity distance-based method, LOF and Random Forests in terms of AUC and processing time, and especially in large data sets. iForest also works well in high dimensional prob-lems which have a large number of irrelevant attributes, and in situations where training set does not contain any anomalies. 1 Text Orca Unknown |
institution |
Open Polar |
collection |
Unknown |
op_collection_id |
ftciteseerx |
language |
English |
description |
Most existing model-based approaches to anomaly de-tection construct a profile of normal instances, then iden-tify instances that do not conform to the normal profile as anomalies. This paper proposes a fundamentally different model-based method that explicitly isolates anomalies in-stead of profiles normal points. To our best knowledge, the concept of isolation has not been explored in current liter-ature. The use of isolation enables the proposed method, iForest, to exploit sub-sampling to an extent that is not fea-sible in existing methods, creating an algorithm which has a linear time complexity with a low constant and a low mem-ory requirement. Our empirical evaluation shows that iFor-est performs favourably to ORCA, a near-linear time com-plexity distance-based method, LOF and Random Forests in terms of AUC and processing time, and especially in large data sets. iForest also works well in high dimensional prob-lems which have a large number of irrelevant attributes, and in situations where training set does not contain any anomalies. 1 |
author2 |
The Pennsylvania State University CiteSeerX Archives |
format |
Text |
author |
Fei Tony Liu Kai Ming Ting Zhi-hua Zhou |
spellingShingle |
Fei Tony Liu Kai Ming Ting Zhi-hua Zhou Isolation Forest |
author_facet |
Fei Tony Liu Kai Ming Ting Zhi-hua Zhou |
author_sort |
Fei Tony Liu |
title |
Isolation Forest |
title_short |
Isolation Forest |
title_full |
Isolation Forest |
title_fullStr |
Isolation Forest |
title_full_unstemmed |
Isolation Forest |
title_sort |
isolation forest |
url |
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.678.3903 http://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/icdm08b.pdf |
genre |
Orca |
genre_facet |
Orca |
op_source |
http://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/icdm08b.pdf |
op_relation |
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.678.3903 http://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/icdm08b.pdf |
op_rights |
Metadata may be used without restrictions as long as the oai identifier remains attached to it. |
_version_ |
1766161477682593792 |