An Experience with the Implementation of a Rule-Based Triggering Recommendation Approach for Mobile Devices

In the current Big Data era, mobile context-aware recommender systems can play a key role to help citizens and tourists to make good decisions. Ideally, these systems should be proactive, able to detect the right moment and place to offer suggestions of a specific type of item or activity to the user. For this purpose, push-based recommender systems can be used, exploiting context rules to decide when a specific type of recommendation should be triggered. However, experiences regarding the implementation of these types of systems are scarce. Motivated by this, in this paper, we describe our design and implementation efforts focusing on the ability to fire suitable recommendations, without user intervention, whenever it is required. In our proposal, the mobile user can activate, deactivate, parametrize, and define rules in an easy way, to obtain a better user personalization. Besides, the recommendation triggering is performed on the mobile device, which allows minimizing the amount of wireless communications and helps to protect the user’s privacy (as context data is evaluated locally on the device, rather than by an external server). We have analyzed several technological options and evaluated the performance and scalability of our proposal, showing its feasibility.