Customized Pricing Recommender System — Simple Implementation and Preliminary Experiments —

Recommender systems suggests items that would be preferred to customers. Here, we propose to add new function, price discounting, to these systems. This new system determines whether it offers discounting for each customer, and thus this price-setting scheme is called price customization. We discuss the benefits that this customized pricing recommender system will bring for both customers and dealers. We propose to realize such systems by combining standard recommendation algorithm and multi-armed bandit approaches. We implemented a simple system and performed preliminary experiments on semi-simulated data.