Interpretable Neural Temporal Point Processes for Modelling Electronic Health Records
This work provides an interpretable method for healthcare professionals to analyze dependencies in medical event sequences, though it is incremental as it builds on existing NTPP and Hawkes process concepts.
The paper tackled the problem of modeling Electronic Health Records as temporal event sequences by addressing the lack of interpretability in existing Neural Temporal Point Process models, resulting in a framework that directly parameterizes event influences and shows superiority in event prediction and influence learning.
Electronic Health Records (EHR) can be represented as temporal sequences that record the events (medical visits) from patients. Neural temporal point process (NTPP) has achieved great success in modeling event sequences that occur in continuous time space. However, due to the black-box nature of neural networks, existing NTPP models fall short in explaining the dependencies between different event types. In this paper, inspired by word2vec and Hawkes process, we propose an interpretable framework inf2vec for event sequence modelling, where the event influences are directly parameterized and can be learned end-to-end. In the experiment, we demonstrate the superiority of our model on event prediction as well as type-type influences learning.