New Paper: Time Matters: Time-Aware LSTMs for Predictive Business Process Monitoring

We are happy to introduce a new publication, which was written in cooperation with the Chair of Digital Industrial Service Systems from FAU: “Time Matters: Time-Aware LSTMs for Predictive Business Process Monitoring”. In this paper An Nguyen, Srijeet Chatterjee, Sven Weinzierl, Leo Schwinn, Martin Matzner, and Björn Eskofier propose a new predictive business process monitoring (PBPM) technique based on time-aware long short-term memory (T-LSTM) cells, which allows for better modelling of time dependencies between events. Furthermore, they introduce cost-sensitive learning to account for the common class imbalance in event logs. The presentation can be watched on Youtube. The code is available on GitHub. A preprint can be found here.

This work was presented at the FIRST INTERNATIONAL WORKSHOP ON LEVERAGING MACHINE LEARNING IN PROCESS MINING co-located with the second International Conference on Process Mining (ICPM).