UCL School of Management is delighted to welcome Kanishka Misra, Rady School of Management, to host a seminar discussing ‘Deep Learning Demand: Integrating Economic Theory with Artificial Intelligence’

Abstract

Current deep learning algorithms provide flexible statistical tools that can uncover complex patterns for a variety of raw data generating process. A critique of such algorithms in Marketing is the lack of underling theory. In this paper, we propose a novel combination of economic theory with machine learning. We consider the context of static discrete choice models with panel data. Firms have access to a large dataset of individual consumer purchases decisions (e.g. which UPC was purchased), detailed  consumer specific characteristics (e.g., demographics), and environment specific characteristics (e.g., price). Our new approach adds choice theory assumptions to deep learning algorithms. In particular, we impose the weak axiom of revealed preference. Our simulations show that imposing economic theory can improve predictions from state of the art machine learning algorithms. We find the magnitude of the improvement depends on the size of the dataset and the complexity of the underlying data generating process. We apply our model to a field panel dataset and show that our model does improve on predictions from computer science models.