Prescriptive analytics through constrained bayesian optimization

Abstract

Prescriptive analytics leverages predictive data mining algorithms to prescribe appropriate changes to alter a predicted outcome of undesired class to a desired one. As an example, based on the conversation of a reformed addict on a message board, prescriptive analytics may predict the intervention required. We develop a novel prescriptive analytics solution by formulating a constrained Bayesian optimization problem to find the smallest change that we need to make on an actionable set of features so that with sufficient confidence an instance can be changed from an undesirable class to the desirable class. We use two public health dataset, multi-year CDC dataset on disease prevalence across the 50 states of USA and alcohol related data from Reddit to demonstrate the usefulness of our results.