ShortFuse: Biomedical Time Series Representations in the Presence of Structured Information

TL;DR

ShortFuse is a deep learning method that improves biomedical time series predictions by integrating structured patient data with temporal features, enhancing accuracy in healthcare applications.

Contribution

It introduces hybrid convolutional and LSTM cells that explicitly incorporate structured covariates into time series modeling, a novel approach in healthcare data analysis.

Findings

Outperforms competing models by 3% in biomedical tasks

Matches or exceeds domain-engineered feature models

Effective in predicting osteoarthritis progression and surgical outcomes

Abstract

In healthcare applications, temporal variables that encode movement, health status and longitudinal patient evolution are often accompanied by rich structured information such as demographics, diagnostics and medical exam data. However, current methods do not jointly optimize over structured covariates and time series in the feature extraction process. We present ShortFuse, a method that boosts the accuracy of deep learning models for time series by explicitly modeling temporal interactions and dependencies with structured covariates. ShortFuse introduces hybrid convolutional and LSTM cells that incorporate the covariates via weights that are shared across the temporal domain. ShortFuse outperforms competing models by 3% on two biomedical applications, forecasting osteoarthritis-related cartilage degeneration and predicting surgical outcomes for cerebral palsy patients, matching or exceeding the accuracy of models that use features engineered by domain experts.