As we gather more information from our processes, we need mathematical models that provide flexibility and accuracy for complex systems. Classical statistical modelling and prediction often fall short in these situations. Bayesian statistical methods, however, are fundamentally flexible and can address these shortcomings. They can be leveraged to improve outcomes from complex systems/relationships made possible with digital transformation.
These advanced methods are not part of typical statistical training received by engineers and scientists. However, it is critical that engineers are aware of these options, so that the data they collect can be leveraged for maximum value. This presentation will cover common situations (e.g., hierarchical models, risk-based stability modelling, staged testing) where innovative modelling provides answers when classic approaches fail.