“…In recent years, DBN models have seen a lot of use in industrial settings due to their characteristics as TS forecasting models and their interpretability, which has changed DBNs into more general use models. They have been applied to stock market forecasting, 3 to ecosystem changes prediction based on climate variations, 4 to topic-sentiment evolution analysis over time, 5 to assess the remaining useful life of structures, 6,7 to monitor aircraft wing cracks evolution over time 8 and to identify abnormal events during cyber security threats, 9 among others. However, in a continuous case, where Gaussianity is typically assumed, DBNs present some drawbacks: DBN models are inherently linear models, and they do not allow the insertion of discrete variables without the introduction of additional constraints.…”