Computationally lightweight classifiers with frequentist bounds on predictions

TL;DR

This paper introduces a fast, resource-efficient classifier based on the Nadaraya-Watson estimator that provides uncertainty bounds, suitable for real-time safety-critical applications.

Contribution

It presents a novel, computationally lightweight classification method with frequentist uncertainty bounds, scalable to large datasets and applicable in resource-constrained environments.

Findings

Achieves over 96% accuracy on benchmark data.

Operates with O(n) and O(log n) complexity, significantly faster than kernel-based methods.

Provides actionable uncertainty bounds for low-confidence prediction flagging.

Abstract

While both classical and neural network classifiers can achieve high accuracy, they fall short on offering uncertainty bounds on their predictions, making them unfit for safety-critical applications. Existing kernel-based classifiers that provide such bounds scale with $\mathcal O (n^{\sim3})$ in time, making them computationally intractable for large datasets. To address this, we propose a novel, computationally efficient classification algorithm based on the Nadaraya-Watson estimator, for whose estimates we derive frequentist uncertainty intervals. We evaluate our classifier on synthetically generated data and on electrocardiographic heartbeat signals from the MIT-BIH Arrhythmia database. We show that the method achieves competitive accuracy $>$\SI{96}{\percent} at $\mathcal O(n)$ and $\mathcal O(\log n)$ operations, while providing actionable uncertainty bounds. These bounds can, e.g., aid in flagging low-confidence predictions, making them suitable for real-time settings with resource constraints, such as diagnostic monitoring or implantable devices.