Time-aware and transition-semantic graph neural networks for interpretable predictive business process monitoring

• GNN for next event prediction using both prefix and full trace inputs. • Time-decay attention enhances focus on temporally relevant events. • Semantic edge embeddings improve modeling of transition dynamics. • Visual interpretability supports trace level model transparency. Predictive Business Process Monitoring (PBPM) aims to forecast future events in ongoing cases based on historical event logs. While Graph Neural Networks (GNNs) are well suited to capture structural dependencies in process data, existing GNN-based PBPM models remain underdeveloped. Most rely either on short prefix subgraphs or global architectures that overlook temporal relevance and transition semantics. We propose a unified, interpretable GNN framework that advances the state of the art along three key axes. First, we compare prefix-based Graph Convolutional Networks(GCNs) and full trace Graph Attention Networks(GATs) to quantify the performance gap between localized and global modeling. Second, we introduce a novel time decay attention mechanism that constructs dynamic, prediction-centered windows, emphasizing temporally relevant history and suppressing noise. Third, we embed transition type semantics into edge features to enable fine grained reasoning over structurally ambiguous traces. Our architecture includes multilevel interpretability modules, offering diverse visualizations of attention behavior. Evaluated on six benchmarks with different decay factors, the proposed models achieve competitive Top- k accuracy and Damerau-Levenshtein scores under a consistent global setup. A separate analysis of the decay coefficient λ confirms its role in adapting attention, with effective values calibrated by temporal structure and sequence depth. By addressing architectural, temporal, and semantic gaps, this work presents a robust, generalizable, and explainable solution for next event prediction in PBPM.