Predictive Process Monitoring (PPM) uses event logs to forecast future events in business processes. Deep learning models have been applied to PPM, but they face a major hurdle: labeled data is often scarce, and real-world event logs tend to have low variability. This limits the effectiveness of these models.
To address this, the arXiv paper "Process-Aware Event Log Augmentation for Process Prediction under Labeled Data Scarcity" proposes a process-aware augmentation technique. The idea is to generate additional training examples by augmenting existing event logs in a way that respects the underlying process structure, rather than using generic data augmentation.
The approach is specifically designed for the labeled data scarcity scenario. By increasing the amount and variability of training data, the method aims to improve the performance of deep learning PPM models. The abstract does not provide experimental results, so the reported effectiveness is not yet known from this announcement.