Multivariate Time-series Similarity Assessment via Unsupervised Representation Learning and Stratified Locality Sensitive Hashing: Application to Early Acute Hypotensive Episode Detection
This work addresses timely prediction of critical events in ICU care, but it is incremental as it builds on existing representation learning and hashing methods applied to a specific medical domain.
The paper tackles the problem of predicting Acute Hypotensive Episodes in ICU patients by eliminating hand-crafted feature engineering from multivariate time-series, using an unsupervised sequence-to-sequence auto-encoder and hashing for similarity assessment, resulting in accurate prediction of upcoming AHE events.
Timely prediction of clinically critical events in Intensive Care Unit (ICU) is important for improving care and survival rate. Most of the existing approaches are based on the application of various classification methods on explicitly extracted statistical features from vital signals. In this work, we propose to eliminate the high cost of engineering hand-crafted features from multivariate time-series of physiologic signals by learning their representation with a sequence-to-sequence auto-encoder. We then propose to hash the learned representations to enable signal similarity assessment for the prediction of critical events. We apply this methodological framework to predict Acute Hypotensive Episodes (AHE) on a large and diverse dataset of vital signal recordings. Experiments demonstrate the ability of the presented framework in accurately predicting an upcoming AHE.