Functional mixture-of-experts for classification

TL;DR

This paper introduces a functional mixture-of-experts model for multiclass classification with univariate functional predictors, employing regularized estimation with interpretability-focused sparsity constraints.

Contribution

It presents a novel ME model with multinomial logistic functions for functional inputs and an EM-Lasso algorithm for regularized estimation.

Findings

Effective on simulated data

Demonstrates interpretability of coefficient functions

Shows promising results on real data

Abstract

We develop a mixtures-of-experts (ME) approach to the multiclass classification where the predictors are univariate functions. It consists of a ME model in which both the gating network and the experts network are constructed upon multinomial logistic activation functions with functional inputs. We perform a regularized maximum likelihood estimation in which the coefficient functions enjoy interpretable sparsity constraints on targeted derivatives. We develop an EM-Lasso like algorithm to compute the regularized MLE and evaluate the proposed approach on simulated and real data.