Probabilistic data-driven predictive models for energy applications.

      Аннотация: The abundance of collected data from physical systems holds the promise of transforming conventional systems into smarter infrastructures that permits the development of credible decisions. These perspectives, taken in the context of recent algorithmic developments for big data, start to address many of the challenges encountered in complex systems. On the other hand, the intertwined uncertainty associated with these systems impose an additional layer of complexity that needs to be addressed properly. Stochastic predictive modeling seeks to quantify the effect of uncertainty on the overall system behavior in order to improve the process of decision making. While fulfilling this promise remains fraught with conceptual, technological, and mathematical challenges, success holds many promises that go to the heart of societal and environmental sustainability.

Ключевые слова: data-driven predictive models, energy applications, conventional systems, infrastructure, development