Integrative Analysis and Imputation of Multiple Data Streams via Deep Gaussian Processes

2025-05-17

AI Summary

Intensive care data present three key challenges: (1) multimodal physiological signals exhibit intrinsic interdependencies yet are commonly modeled in isolation; (2) observations are irregularly sampled with clinically meaningful timing; and (3) missing values are pervasive, while existing imputation methods neglect temporal structure and uncertainty quantification. To address these, we propose the first unified modeling framework integrating deep Gaussian processes (DGPs) with stochastic imputation. Our approach jointly captures longitudinal dynamics and cross-sectional, cross-modal dependencies to enable coherent imputation of asynchronous, multivariate time series, while producing well-calibrated uncertainty estimates. Evaluated on real-world clinical datasets, our method significantly outperforms state-of-the-art baselines—including MICE, last-observation-carried-forward, and univariate Gaussian processes—achieving new state-of-the-art performance in both imputation accuracy and uncertainty calibration.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Intelligent Robots: State Estimation

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Healthcare data, particularly in critical care settings, presents three key challenges for analysis. First, physiological measurements come from different sources but are inherently related. Yet, traditional methods often treat each measurement type independently, losing valuable information about their relationships. Second, clinical measurements are collected at irregular intervals, and these sampling times can carry clinical meaning. Finally, the prevalence of missing values. Whilst several imputation methods exist to tackle this common problem, they often fail to address the temporal nature of the data or provide estimates of uncertainty in their predictions. We propose using deep Gaussian process emulation with stochastic imputation, a methodology initially conceived to deal with computationally expensive models and uncertainty quantification, to solve the problem of handling missing values that naturally occur in critical care data. This method leverages longitudinal and cross-sectional information and provides uncertainty estimation for the imputed values. Our evaluation of a clinical dataset shows that the proposed method performs better than conventional methods, such as multiple imputations with chained equations (MICE), last-known value imputation, and individually fitted Gaussian Processes (GPs).

Problem

• Handling missing values in irregularly sampled clinical data
• Capturing relationships between diverse physiological measurements
• Providing uncertainty estimates for imputed healthcare data values

Innovation

• Deep Gaussian processes for data imputation
• Handles irregular and missing clinical measurements
• Provides uncertainty estimates for imputed values