Faisal Awartani (Ph.D.)
The idea of predictive analytics, either using statistical modeling (time-series analysis) or machine learning (convolutional neural networks), makes a basic premise at the epistemological level, which is a certain structure repeats itself over time-space. That is there is an underlying ordinance within a given social or physical context that is time independent. If you want to extrapolate this concept to the universe, this premise translates to assuming the existence of a set of underlying time-independent rules that runs the universe. In computer science terminology, these rules could metaphorically represent the software that runs the hardware of the universe. If we assume there is no such time independent ordinance that governs the universe physical infrastructure, then the idea of predictive analytics loses its value. Especially when we assume that each day carries with it new experiences with no underlying time-independent structure. This implies that the universe manifests itself in a random manner with no underlying construct that can be expressed through a mathematical or neural network model.