Functional mixture-of-experts for classification
This work addresses classification problems in domains with functional data, such as time series or signal processing, but is incremental as it adapts existing mixture-of-experts methods to functional inputs.
The authors tackled multiclass classification with functional predictors by developing a mixture-of-experts model using multinomial logistic activation functions and sparsity constraints, achieving improved performance on simulated and real data.
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.