New Developments in Unsupervised Outlier Detection by Xiaochun Wang & Xiali Wang & Mitch Wilkes

Author:Xiaochun Wang & Xiali Wang & Mitch Wilkes
Language: eng
Format: epub
ISBN: 9789811595196
Publisher: Springer Singapore

For detecting top 6 outliers, COF misses B and C, LDOF misses E and F, RBDA misses C, and PBOD misses all. The good news is that DB, DB-Max, LOF, INFLO, MST + LOG, and our method detect all six outliers correctly but with different rankings. If the sparse cluster (i.e., the one consisting of 7 data points) is also regarded as an outlying group, k is set to 8. For this situation, both DB-Max and LOF miss 1, COF misses 6, INFLO misses 2, LDOF misses 8, RBDA misses 6, PBOD misses 6, LOCI misses 8, while DB, MST + LOF, and our outlier detector get all right, as shown in Fig. 5.10.

Fig. 5.10Outlier detecting results for Synthetic Dataset 3 for k = 8, reprinted from Ref. [38], copyright 2015, with permission from Elsevier

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.