Interaction commerce, a technological architecture focused on recommender system

Abstract

This paper aims at introducing a type of social commerce architecture to which the name Interaction Commerce has been given. First, a global description of the main macro-components forming the structure of this architecture is provided. Such components also take care of managing e-commerce activities and social relationships within the architecture. Second, the focus is set on the analysis of the single components that are key to the social aspects of the architecture. A special chapter is then entirely focussed on a topic that is considered extremely important by the entire research community, i.e., recommender systems. After providing a general introduction on the topic, the two most common recommendation approaches are analyzed and compared. These are the content-based approach and the collaborative filtering approach. The analysis has shown how all recommender systems are threatened by the cold-start problem. Studying recommender systems has allowed for their implementation in the architecture, which now has a new "social" approach that is able to solve the new user cold-start problem. An architecture prototype was developed and tested in order to be validated.