Dimensionality-Aware Outlier Detection
Document Type
Conference Proceeding
Publication Date
1-1-2024
Abstract
We present a nonparametric method for outlier detection that takes full account of local variations in intrinsic dimensionality within the dataset. Using the theory of Local Intrinsic Dimensionality (LID), our 'dimensionality-aware' outlier detection method, DAO, is derived as an estimator of an asymptotic local expected density ratio involving the query point and a close neighbor drawn at random. The dimensionality-aware behavior of DAO is due to its use of local estimation of LID values in a theoretically-justified way. Through comprehensive experimentation on more than 800 synthetic and real datasets, we show that DAO significantly outperforms three popular and important benchmark outlier detection methods: Local Outlier Factor (LOF), Simplified LOF, and kNN.
Identifier
85182580186 (Scopus)
ISBN
[9781611978032]
Publication Title
Proceedings of the 2024 SIAM International Conference on Data Mining, SDM 2024
First Page
652
Last Page
660
Fund Ref
Danmarks Frie Forskningsfond
Recommended Citation
Anderberg, Alastair; Bailey, James; Campello, Ricardo J.G.B.; Houle, Michael E.; Marques, Henrique O.; Radovanović, Miloš; and Zimek, Arthur, "Dimensionality-Aware Outlier Detection" (2024). Faculty Publications. 1112.
https://digitalcommons.njit.edu/fac_pubs/1112
