Varying Coefficient Linear Discriminant Analysis for Dynamic Data

TL;DR

This paper introduces a varying coefficient LDA model for dynamic data, employing B-spline based least squares estimation, which improves computational efficiency and achieves near-optimal error bounds, validated through experiments.

Contribution

It proposes a novel B-spline based estimation method for varying coefficient LDA, enhancing computational efficiency and theoretical error bounds for dynamic data classification.

Findings

The method outperforms existing dynamic classification techniques.

Estimation error bounds are optimal or near optimal in different regimes.

Numerical experiments confirm the method's superiority.

Abstract

Linear discriminant analysis (LDA) is an important classification tool in statistics and machine learning. This paper investigates the varying coefficient LDA model for dynamic data, with Bayes' discriminant direction being a function of some exposure variable to address the heterogeneity. We propose a new least-square estimation method based on the B-spline approximation. The data-driven discriminant procedure is more computationally efficient than the dynamic linear programming rule \citep{jiang2020dynamic}. We also establish the convergence rates for the corresponding estimation error bound and the excess misclassification risk. The estimation error in $L_2$ distance is optimal for the low-dimensional regime and is near optimal for the high-dimensional regime. Numerical experiments on synthetic data and real data both corroborate the superiority of our proposed classification method.