A model-free subdata selection method for classification

TL;DR

This paper introduces a model-free subdata selection method called PED, which uses decision trees and random forests to efficiently select representative data for classification, improving accuracy over existing methods.

Contribution

The paper proposes a novel model-free subdata selection approach for classification that does not rely on underlying model assumptions, enhancing robustness and accuracy.

Findings

PED subdata results in smaller Gini index than uniform sampling.

PED subdata achieves higher classification accuracy in simulations.

Method is effective for multiple classes and various predictor types.

Abstract

Subdata selection is a study of methods that select a small representative sample of the big data, the analysis of which is fast and statistically efficient. The existing subdata selection methods assume that the big data can be reasonably modeled using an underlying model, such as a (multinomial) logistic regression for classification problems. These methods work extremely well when the underlying modeling assumption is correct but often yield poor results otherwise. In this paper, we propose a model-free subdata selection method for classification problems, and the resulting subdata is called PED subdata. The PED subdata uses decision trees to find a partition of the data, followed by selecting an appropriate sample from each component of the partition. Random forests are used for analyzing the selected subdata. Our method can be employed for a general number of classes in the response and for both categorical and continuous predictors. We show analytically that the PED subdata results in a smaller Gini than a uniform subdata. Further, we demonstrate that the PED subdata has higher classification accuracy than other competing methods through extensive simulated and real datasets.