Sep 2026· Statistical analysis and data mining· Vol 19· 0 citations· 24 references
Computer Science
TL;DR
A robust loss function tailored to handle outliers effectively applied within a neural network based on basis expansion is introduced, and a novel functional normalization technique that normalizes the input function values of hidden layers with respect to time points is proposed.
Abstract
This paper aims to develop a robust nonlinear model utilizing neural networks for function‐on‐function regression problems. The proposed framework is designed to enhance robustness in regression modeling by minimizing the impact of outliers in the response variable during model training. To achieve this, we introduce a robust loss function tailored to handle outliers effectively applied within a neural network based on basis expansion, and propose a novel functional normalization technique that normalizes the input function values of hidden layers with respect to time points. This approach demonstrates faster convergence as compared to existing neural network models and is less sensitive to hyperparameter settings. It also achieves high performance regression results even with outliers in the response variable. The effectiveness and superiority of the proposed method are validated through extensive simulation studies and real‐world applications using COVID‐19 excess mortality data.
Feature selection in neural networks remains a challenging problem, particularly in the presence of noisy or contaminated data. LassoNet is a recent approach that addresses this issue by combining neural networks with hierarchical sparsity constraints, enabling simultaneous prediction and variable selection. However, i...
D. De Canditiis, I. De Feis, P. Stolfi· arXiv.org· 0 citations
In this paper, we investigate robust model averaging for high-dimensional data with outlier contamination. Most existing model averaging methods rely on ordinary least squares (OLS) estimation and are therefore highly sensitive to outliers. Moreover, these methods are generally not designed to accommodate high-dimensio...
An improved SNN using virtual space (VSNN) was introduced, which has unique advantages in online training or equidistant-sampling problems and can effectively supplement the SNN algorithm system, promoting the bionic research of artificial neural networks.
This work introduces total sensitivity kernels (TSKs), a method based on families of weighted ANOVA kernels that learn and adapt to this multivariable structure, and establishes consistency of a finite-data formulation based on minimum-norm interpolation.
John E. Darges, Laura Weidensager· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.