Fuzzy Personalization of Mobile Apps: A Case Study from mHealth Domain

The growing interest in wearable devices has stimulated the development of mHealth applications: users can be monitored at different levels of granularity and their data can be exploited for recommendations about different aspects of their conditions, i.e., physical, psychological, and social. To this aim, recommendation systems should be able to profile patients to suggest them the most proper actions to promote effective behavior changes. This paper presents a solution to this challenging research topic implemented in an Android app, based on the adoption of fuzzy logic to cluster users according to quantitative and qualitative variables about their physical and psychological well-being. Four classes have been obtained from the two models developed, in accordance with previous experiments. The final aim of user profiling is promoting group physical activity among users characterized by similar behaviors.