Linear Algorithms for Robust and Scalable Nonparametric Multiclass Probability Estimation

TL;DR

This paper introduces two linear-time algorithms for multiclass probability estimation using weighted SVMs, improving computational efficiency and accuracy while maintaining consistency and robustness.

Contribution

It proposes baseline and One-vs-All learning schemes that enhance weighted SVMs for multiclass probability estimation, with linear complexity and improved accuracy.

Findings

Baseline learning has linear complexity in K.

OVA learning achieves the highest estimation accuracy.

Estimators are distribution-free and consistent.

Abstract

Multiclass probability estimation is the problem of estimating conditional probabilities of a data point belonging to a class given its covariate information. It has broad applications in statistical analysis and data science. Recently a class of weighted Support Vector Machines (wSVMs) has been developed to estimate class probabilities through ensemble learning for $K$-class problems (Wu, Zhang and Liu, 2010; Wang, Zhang and Wu, 2019), where $K$ is the number of classes. The estimators are robust and achieve high accuracy for probability estimation, but their learning is implemented through pairwise coupling, which demands polynomial time in $K$. In this paper, we propose two new learning schemes, the baseline learning and the One-vs-All (OVA) learning, to further improve wSVMs in terms of computational efficiency and estimation accuracy. In particular, the baseline learning has optimal computational complexity in the sense that it is linear in $K$. Though not being most efficient in computation, the OVA offers the best estimation accuracy among all the procedures under comparison. The resulting estimators are distribution-free and shown to be consistent. We further conduct extensive numerical experiments to demonstrate finite sample performance.