Average Localised Proximity: A new data descriptor with good default one-class classification performance

TL;DR

This paper introduces Average Localised Proximity (ALP), a new data descriptor for one-class classification that outperforms existing methods like Isolation Forest and is a strong default choice due to its optimal hyperparameters.

Contribution

The paper proposes ALP, a novel data descriptor for one-class classification, and systematically determines optimal default hyperparameters across diverse real-world datasets.

Findings

ALP outperforms Isolation Forest in one-class classification tasks.

ALP shows weak but consistent improvement over SVM.

Optimal default hyperparameters enhance ease of use and performance.

Abstract

One-class classification is a challenging subfield of machine learning in which so-called data descriptors are used to predict membership of a class based solely on positive examples of that class, and no counter-examples. A number of data descriptors that have been shown to perform well in previous studies of one-class classification, like the Support Vector Machine (SVM), require setting one or more hyperparameters. There has been no systematic attempt to date to determine optimal default values for these hyperparameters, which limits their ease of use, especially in comparison with hyperparameter-free proposals like the Isolation Forest (IF). We address this issue by determining optimal default hyperparameter values across a collection of 246 one-class classification problems derived from 50 different real-world datasets. In addition, we propose a new data descriptor, Average Localised Proximity (ALP) to address certain issues with existing approaches based on nearest neighbour distances. Finally, we evaluate classification performance using a leave-one-dataset-out procedure, and find strong evidence that ALP outperforms IF and a number of other data descriptors, as well as weak evidence that it outperforms SVM, making ALP a good default choice.