Proactive Mobile CARS in Action: A First Step Towards Making Sense of Context Rules

Abstract

Recommender systems play a key role towards the development of personalization services, as they are able to provide suggestions about specific types of items (points of interest in a city, restaurants, hotels, etc.) that a particular user may find relevant, based on his/her preferences. In recent years, it has been argued that it is important to consider the context of the user (e.g., his/her location, the time of the day, etc.) to offer suitable recommendations to mobile users, which has given rise to the so-called Context-Aware Recommender Systems (CARS). Moreover, to facilitate users the access to relevant information and minimize the required interaction effort, they should receive the recommendations proactively, without the need to explicitly ask for a specific type of item.However, more research is needed to determine the impact of different context attributes on specific scenarios as well as the conditions under which recommendations of some types of items should be automatically activated. In this paper, we focus on the problem of recommendation triggering, describe some use case scenarios, and present context attributes and rules that can be defined to initiate several types of recommendations appropriate for those scenarios. For illustration, we formulate some examples of conditions as SWRL-like rules defined over the Semantic Sensor Network (SSN) ontology.