Prototype selection for interpretable classification

TL;DR

This paper introduces a prototype selection method for classification that identifies a minimal, representative subset of samples to enhance interpretability and efficiency in large data sets.

Contribution

It proposes a set cover-based approach for selecting prototypes, emphasizing interpretability and sparsity in the sample space rather than just variables.

Findings

The method effectively reduces data complexity while maintaining classification accuracy.

Prototypes provide meaningful interpretative summaries of data.

The approach can be used both for interpretability and as a classifier.

Abstract

Prototype methods seek a minimal subset of samples that can serve as a distillation or condensed view of a data set. As the size of modern data sets grows, being able to present a domain specialist with a short list of "representative" samples chosen from the data set is of increasing interpretative value. While much recent statistical research has been focused on producing sparse-in-the-variables methods, this paper aims at achieving sparsity in the samples. We discuss a method for selecting prototypes in the classification setting (in which the samples fall into known discrete categories). Our method of focus is derived from three basic properties that we believe a good prototype set should satisfy. This intuition is translated into a set cover optimization problem, which we solve approximately using standard approaches. While prototype selection is usually viewed as purely a means toward building an efficient classifier, in this paper we emphasize the inherent value of having a set of prototypical elements. That said, by using the nearest-neighbor rule on the set of prototypes, we can of course discuss our method as a classifier as well.