SWoTTeD: An Extension of Tensor Decomposition to Temporal Phenotyping

TL;DR

SWoTTeD is a novel tensor decomposition method designed to uncover meaningful temporal phenotypes in complex time-series data like EHR, improving interpretability and maintaining high accuracy.

Contribution

It introduces SWoTTeD, a new approach that extends tensor decomposition with temporal constraints for better phenotyping in healthcare data.

Findings

Achieves comparable reconstruction accuracy to state-of-the-art models.

Extracts interpretable temporal phenotypes relevant for clinicians.

Validated on synthetic and real-world datasets, including hospital data.

Abstract

Tensor decomposition has recently been gaining attention in the machine learning community for the analysis of individual traces, such as Electronic Health Records (EHR). However, this task becomes significantly more difficult when the data follows complex temporal patterns. This paper introduces the notion of a temporal phenotype as an arrangement of features over time and it proposes SWoTTeD (Sliding Window for Temporal Tensor Decomposition), a novel method to discover hidden temporal patterns. SWoTTeD integrates several constraints and regularizations to enhance the interpretability of the extracted phenotypes. We validate our proposal using both synthetic and real-world datasets, and we present an original usecase using data from the Greater Paris University Hospital. The results show that SWoTTeD achieves at least as accurate reconstruction as recent state-of-the-art tensor decomposition models, and extracts temporal phenotypes that are meaningful for clinicians.