Decision Forest: A Nonparametric Approach to Modeling Irrational Choice

Customer behavior is often assumed to follow weak rationality, which implies that adding a product an assortment will not increase the choice probability of another in assortment. However, increasing amount research has revealed customers are necessarily rational when making decisions. In this paper, we propose new nonparametric model relaxes assumption and can wider range customer behavior, such as decoy effects between products. model, each type associated with binary decision tree, represents process for purchase based on checking existence specific products Together distribution over types, show resulting model—a forest—is able represent any including models inconsistent rationality. We theoretically characterize depth forest needed fit data set historical assortments prove high probability, whose scales logarithmically number sufficient most sets. also two practical algorithms—one column generation one random sampling—for estimating from data. Using synthetic real transaction exhibiting nonrational outperforms both benchmark out-of-sample predictive ability. This paper was accepted by Chung Piaw Teo, optimization.