The Data Replication Method for the Classification with Reject Option

TL;DR

This paper introduces a novel data replication method for classification with reject options, transforming the problem into standard classification tasks and extending it to multiclass ordinal data, validated through experiments.

Contribution

It adapts a paradigm for ordinal data classification to handle reject options, integrating it with SVMs and neural networks, and extends it to multiclass scenarios.

Findings

Effective reduction of reject option classification to standard problems

Successful implementation with SVMs and neural networks

Validated on synthetic and real datasets

Abstract

Classification is one of the most important tasks of machine learning. Although the most well studied model is the two-class problem, in many scenarios there is the opportunity to label critical items for manual revision, instead of trying to automatically classify every item. In this paper we adapt a paradigm initially proposed for the classification of ordinal data to address the classification problem with reject option. The technique reduces the problem of classifying with reject option to the standard two-class problem. The introduced method is then mapped into support vector machines and neural networks. Finally, the framework is extended to multiclass ordinal data with reject option. An experimental study with synthetic and real data sets, verifies the usefulness of the proposed approach.