Monash University
Browse
- No file added yet -

On normalization and algorithm selection for unsupervised outlier detection

Download (2.97 MB)
journal contribution
posted on 2022-11-09, 04:40 authored by Sevvandi Kandanaarachchi, Mario A Munoz, Rob J Hyndman, Kate Smith-Miles
This paper demonstrates that the performance of various outlier detection methods depends sensitively on both the data normalization schemes employed, as well as characteristics of the datasets. Recasting the challenge of understanding these dependencies as an algorithm selection problem, we perform the first instance space analysis of outlier detection methods. Such analysis enables the strengths and weaknesses of unsupervised outlier detection methods to be visualized and insights gained into which method and normalization scheme should be selected to obtain the most likely best performance for a given dataset.

History

Classification-JEL

--

Creation date

2018-09-13

Working Paper Series Number

16/18

Length

33

File-Format

application/pdf

Handle

RePEc:msh:ebswps:2018-16

Usage metrics

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC