The Application of xLSTM in Predicting the Next Event in Process Logs

The development of modern information technology has generated complex workflow data in various fields, and process mining on these data can assist managers in better managing and supporting business operations. As an important branch of process mining, predictive process monitoring provides prerequisites for downstream tasks such as risk assessment and decision-making by predicting the future behavior of processes. However, due to the complexity and unstructured nature of business processes, traditional process prediction methods face certain limitations in practical applications. In recent years, significant progress has been made in the application of deep learning in process prediction. This paper introduces an emerging deep learning model, XLSTM (eXtreme Long Short-Term Memory), which demonstrates excellent memory capacity and robustness in handling complex process data spanning long time intervals through optimized gating mechanisms and memory cell design. The purpose of this paper is to explore the application potential of XLSTM in process prediction and quantify its performance in different prediction tasks, in order to evaluate its accuracy and advantages compared to traditional methods. Through experimental verification, this paper validates the performance of XLSTM in handling complex business processes, providing a new direction for future research in process prediction.