Temporal Stability in Predictive Process Monitoring
This addresses the need for reliable decision-making in business processes by focusing on prediction stability, representing an incremental improvement over traditional accuracy-only methods.
The paper tackled the problem of ensuring stable predictions over time in predictive process monitoring, finding that XGBoost and LSTM methods show the highest temporal stability, and demonstrated that hyperparameter optimization and smoothing techniques can improve stability with a slight accuracy trade-off.
Predictive process monitoring is concerned with the analysis of events produced during the execution of a business process in order to predict as early as possible the final outcome of an ongoing case. Traditionally, predictive process monitoring methods are optimized with respect to accuracy. However, in environments where users make decisions and take actions in response to the predictions they receive, it is equally important to optimize the stability of the successive predictions made for each case. To this end, this paper defines a notion of temporal stability for binary classification tasks in predictive process monitoring and evaluates existing methods with respect to both temporal stability and accuracy. We find that methods based on XGBoost and LSTM neural networks exhibit the highest temporal stability. We then show that temporal stability can be enhanced by hyperparameter-optimizing random forests and XGBoost classifiers with respect to inter-run stability. Finally, we show that time series smoothing techniques can further enhance temporal stability at the expense of slightly lower accuracy.