MAGIC: Multi-task Gaussian process for joint imputation and classification in healthcare time series
This addresses the problem of improving patient diagnosis and management in healthcare by enabling more informed treatment decisions, though it is an incremental advancement over traditional two-step methods.
The authors tackled the challenges of time misalignment and data sparsity in healthcare time series by proposing MAGIC, a unified framework for joint imputation and classification, which achieved superior predictive accuracy in applications like predicting post-traumatic headache improvement and in-hospital mortality.
Time series analysis has emerged as an important tool for improving patient diagnosis and management in healthcare applications. However, these applications commonly face two critical challenges: time misalignment and data sparsity. Traditional approaches address these issues through a two-step process of imputation followed by prediction. We propose MAGIC (Multi-tAsk Gaussian Process for Imputation and Classification), a novel unified framework that simultaneously performs class-informed missing value imputation and label prediction within a hierarchical multi-task Gaussian process coupled with functional logistic regression. To handle intractable likelihood components, MAGIC employs Taylor expansion approximations with bounded error analysis, and parameter estimation is performed using EM algorithm with block coordinate optimization supported by convergence analysis. We validate MAGIC through two healthcare applications: prediction of post-traumatic headache improvement following mild traumatic brain injury and prediction of in-hospital mortality within 48 hours after ICU admission. In both applications, MAGIC achieves superior predictive accuracy compared to existing methods. The ability to generate real-time and accurate predictions with limited samples facilitates early clinical assessment and treatment planning, enabling healthcare providers to make more informed treatment decisions.