Deep Neural Mixed Models for Nonlinear Longitudinal Prediction with Subject-Specific Heterogeneity

Abstract

Longitudinal health data often exhibit within-subject correlation, subject-specific heterogeneity, and complex nonlinear predictor--response relationships that may not be adequately captured by conventional mixed models. We propose a Deep Generalized Neural Network Mixed Model (DGNMM) that extends the GLMM framework by using a deep neural network to learn nonlinear population-level relationships while incorporating directly trainable, regularized subject-specific effects to capture individual heterogeneity. The model is estimated using prediction-oriented stochastic optimization with separate regularization of the neural-network parameters and subject-specific effects. This formulation allows us to investigate when increased nonlinear representation capacity provides additional predictive benefit in the presence of subject-specific heterogeneity. A simulation study varies population-level nonlinearity and subject-specific heterogeneity and compares DGNMM with GLMM, artificial neural network (ANN), Maity--Pal (MP), and GNMM alternatives for continuous and binary longitudinal outcomes. In the real-data application, DGNMM achieved the lowest MSPE for continuous depression and anxiety prediction, while alternative models remained competitive for binary outcomes. Exploratory subject-level analyses showed heterogeneous predictive gains across individuals, but the magnitude of the estimated subject-specific effects was not significantly associated with predictive gain. Overall, increased nonlinear representation capacity was most beneficial for continuous prediction under nonlinear response structures, while greater subject-specific heterogeneity did not systematically yield larger predictive gains.

The authors’ abstract, as arXiv lists it. arXiv’s listings, abstracts included, are free for anyone to reuse.

More in Methodology

• arXiv preprint · 12 October 2026

  Distributional Regression for Arbitrary Response Variables Using ReLU Neural Networks

  Junpeng Ren, Carlos Misael Madrid Padilla and Oscar Hernan Madrid Padilla
• arXiv preprint · 11 October 2026

  Adaptive Robust Confidence Sets in High Dimensions

  Ilias Diakonikolas, Chao Gao, Daniel M. Kane and 1 other
• arXiv preprint · 12 October 2026

  When Sample-Splitting Choices Matter in TMLE and AIPW: Cross-Fitting, Targeting, and Repetition

  M. Ehsan Karim and Rainie Fu
• arXiv preprint · 11 October 2026

  Small area learning

  Katarzyna Reluga, Nicola Salvati, Setareh Ranjbar and 2 others
• arXiv preprint · 12 October 2026

  Online Individualized Treatment Selection under Partial Feedback: False Discovery Control and Minimax Optimality

  Runzhe Gu, Zinan Zhao, Wenguang Sun and 1 other

All new papers in Methodology