Business analytics has evolved from being a novelty used by a select few to an accepted facet of conducting business. Recommender systems form a critical component of the business analytics toolkit and, by enabling firms to effectively target customers with products and services, are helping alter the e-commerce landscape. A variety of methods exist for providing recommendations, with collaborative filtering, matrix factorization, and association-rule-based methods being the most common. In this paper, we propose a method to improve the quality of recommendations made using association rules. This is accomplished by combining rules when possible and stands apart from existing rule-combination methods in that it is strongly grounded in probability theory. Combining rules requires the identification of the best combination of rules from the many combinations that might exist, and we use a maximum-likelihood framework to compare alternative combinations. Because it is impractical to apply the maximum likelihood framework directly in real time, we show that this problem can equivalently be represented as a set partitioning problem by translating it into an information theoretic contextÑthe best solution corresponds to the set of rules that leads to the highest sum of mutual information associated with the rules. Through a variety of experiments that evaluate the quality of recommendations made using the proposed approach, we show that (i) a greedy heuristic used to solve the maximum likelihood estimation problem is very effective, providing results comparable to those from using the optimal set partitioning solution; (ii) the recommendations made by our approach are more accurate than those made by a variety of state-of-the-art benchmarks, including collaborative filtering and matrix factorization; and (iii) the recommendations can be made in a fraction of a second on a desktop computer, making it practical to use in real-world applications.
How do recommender systems affect prices and profits of firms under competition? To explore this question, we model the strategic behavior of customers who make repeated purchases at two competing firms: one that provides personalized recommendations and another that does not. When a customer intends to purchase a product, she obtains recommendations from the personalizing firm and uses this recommendation to eventually purchase from one of the firms. The personalizing firm profiles the customer (based on past purchases) to recommend products. Hence, if a customer purchases less frequently from the personalizing firm, the recommendations made to her become less relevant. While considering the impact on the quality of recommendations received, a customer must balance two opposing forces: (1) the lower price charged by the non-personalizing firm, and (2) an additional fit cost incurred when purchasing from the non-personalizing firm and the increased cost due to recommendations of reduced quality in the future. An outcome of the analysis is that the customers should distribute their purchases across both firms to maximize surplus over a planning horizon. Anticipating this response, the firms simultaneously choose prices. We study the sensitivity of the equilibrium prices and profits of the firms with respect to the effectiveness of the recommender system and the profile deterioration rate. We also analyze some interesting variants of the base model in order to study how its key results could be influenced. One of the key takeaways of this research is that the recommender system can influence the price and profit of not only the personalizing firm but also the non-personalizing firm. > >