LoRAS: An oversampling approach for imbalanced datasets

TL;DR

LoRAS is a novel oversampling method for imbalanced datasets that improves both F1-Score and Balanced accuracy by better approximating the minority class data manifold, outperforming SMOTE and its extensions.

Contribution

This paper introduces LoRAS, a new oversampling technique that overcomes SMOTE's over-generalization issue by using localized affine shadowsampling to better model the minority class.

Findings

LoRAS outperforms SMOTE and its extensions in F1-Score and Balanced accuracy.

LoRAS provides a more accurate estimate of the local minority class data distribution.

Experimental results on 14 datasets demonstrate the effectiveness of LoRAS.

Abstract

The Synthetic Minority Oversampling TEchnique (SMOTE) is widely-used for the analysis of imbalanced datasets. It is known that SMOTE frequently over-generalizes the minority class, leading to misclassifications for the majority class, and effecting the overall balance of the model. In this article, we present an approach that overcomes this limitation of SMOTE, employing Localized Random Affine Shadowsampling (LoRAS) to oversample from an approximated data manifold of the minority class. We benchmarked our algorithm with 14 publicly available imbalanced datasets using three different Machine Learning (ML) algorithms and compared the performance of LoRAS, SMOTE and several SMOTE extensions that share the concept of using convex combinations of minority class data points for oversampling with LoRAS. We observed that LoRAS, on average generates better ML models in terms of F1-Score and Balanced accuracy. Another key observation is that while most of the extensions of SMOTE we have tested, improve the F1-Score with respect to SMOTE on an average, they compromise on the Balanced accuracy of a classification model. LoRAS on the contrary, improves both F1 Score and the Balanced accuracy thus produces better classification models. Moreover, to explain the success of the algorithm, we have constructed a mathematical framework to prove that LoRAS oversampling technique provides a better estimate for the mean of the underlying local data distribution of the minority class data space.