Cargando…

Fast Outlier Detection Using a Grid-Based Algorithm

As one of data mining techniques, outlier detection aims to discover outlying observations that deviate substantially from the reminder of the data. Recently, the Local Outlier Factor (LOF) algorithm has been successfully applied to outlier detection. However, due to the computational complexity of...

Descripción completa

Detalles Bibliográficos
Autores principales: Lee, Jihwan, Cho, Nam-Wook
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Public Library of Science 2016
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5104428/
https://www.ncbi.nlm.nih.gov/pubmed/27832163
http://dx.doi.org/10.1371/journal.pone.0165972
_version_ 1782466745619972096
author Lee, Jihwan
Cho, Nam-Wook
author_facet Lee, Jihwan
Cho, Nam-Wook
author_sort Lee, Jihwan
collection PubMed
description As one of data mining techniques, outlier detection aims to discover outlying observations that deviate substantially from the reminder of the data. Recently, the Local Outlier Factor (LOF) algorithm has been successfully applied to outlier detection. However, due to the computational complexity of the LOF algorithm, its application to large data with high dimension has been limited. The aim of this paper is to propose grid-based algorithm that reduces the computation time required by the LOF algorithm to determine the k-nearest neighbors. The algorithm divides the data spaces in to a smaller number of regions, called as a “grid”, and calculates the LOF value of each grid. To examine the effectiveness of the proposed method, several experiments incorporating different parameters were conducted. The proposed method demonstrated a significant computation time reduction with predictable and acceptable trade-off errors. Then, the proposed methodology was successfully applied to real database transaction logs of Korea Atomic Energy Research Institute. As a result, we show that for a very large dataset, the grid-LOF can be considered as an acceptable approximation for the original LOF. Moreover, it can also be effectively used for real-time outlier detection.
format Online
Article
Text
id pubmed-5104428
institution National Center for Biotechnology Information
language English
publishDate 2016
publisher Public Library of Science
record_format MEDLINE/PubMed
spelling pubmed-51044282016-12-08 Fast Outlier Detection Using a Grid-Based Algorithm Lee, Jihwan Cho, Nam-Wook PLoS One Research Article As one of data mining techniques, outlier detection aims to discover outlying observations that deviate substantially from the reminder of the data. Recently, the Local Outlier Factor (LOF) algorithm has been successfully applied to outlier detection. However, due to the computational complexity of the LOF algorithm, its application to large data with high dimension has been limited. The aim of this paper is to propose grid-based algorithm that reduces the computation time required by the LOF algorithm to determine the k-nearest neighbors. The algorithm divides the data spaces in to a smaller number of regions, called as a “grid”, and calculates the LOF value of each grid. To examine the effectiveness of the proposed method, several experiments incorporating different parameters were conducted. The proposed method demonstrated a significant computation time reduction with predictable and acceptable trade-off errors. Then, the proposed methodology was successfully applied to real database transaction logs of Korea Atomic Energy Research Institute. As a result, we show that for a very large dataset, the grid-LOF can be considered as an acceptable approximation for the original LOF. Moreover, it can also be effectively used for real-time outlier detection. Public Library of Science 2016-11-10 /pmc/articles/PMC5104428/ /pubmed/27832163 http://dx.doi.org/10.1371/journal.pone.0165972 Text en © 2016 Lee, Cho http://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/) , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
spellingShingle Research Article
Lee, Jihwan
Cho, Nam-Wook
Fast Outlier Detection Using a Grid-Based Algorithm
title Fast Outlier Detection Using a Grid-Based Algorithm
title_full Fast Outlier Detection Using a Grid-Based Algorithm
title_fullStr Fast Outlier Detection Using a Grid-Based Algorithm
title_full_unstemmed Fast Outlier Detection Using a Grid-Based Algorithm
title_short Fast Outlier Detection Using a Grid-Based Algorithm
title_sort fast outlier detection using a grid-based algorithm
topic Research Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5104428/
https://www.ncbi.nlm.nih.gov/pubmed/27832163
http://dx.doi.org/10.1371/journal.pone.0165972
work_keys_str_mv AT leejihwan fastoutlierdetectionusingagridbasedalgorithm
AT chonamwook fastoutlierdetectionusingagridbasedalgorithm