Multistep prediction of dynamic uncertainty under limited data

Engineering systems are growing in complexity, requiring increasingly intelligent and flexible methods to account for predict uncertainties service. This paper presents a framework dynamic uncertainty prediction under limited data (UPLD). Spatial geometry is incorporated with LSTM networks enable real-time multistep of quantitative qualitative over time. Validation achieved through two case studies. Results demonstrate robust trends parallel determination geometric symmetry at each time unit. Future work recommended explore alternative network architectures suited scenarios.