Time-Aware Transformer-based Network for Clinical Notes Series Prediction
Time-Aware Transformer-based Network for Clinical Notes Series Prediction
Dongyu Zhang,Jidapa Thadajarassiri,Cansu Sen,Elke A. Rundensteiner
2020 · DBLP: conf/mlhc/ZhangTSR20
Machine Learning in Health Care · 42 Citations
TLDR
A time-aware transformer-based hierarchical architecture, which is called F lexible T ime-aware L STM Trans former ( FTL-Trans), for classifying a patient’s health state based on her series of clinical notes, successfully captures the multi-level sequential structure of clinical note series.