Recent efforts regarding mobile social networks aim at connecting people in smart environments considering not only their social behavior but also their context. In this perspective, this work presents a model, named SSN, based on the concept of Spontaneous Social Network, which allows the interaction of people anywhere and anytime without the need of any preexisting relation. The main scientific contribution of the SSN model is the possibility of creating social communities based on a combination of multiple contexts, including location, profile and data obtained from external online social networks. In this way, SSN focus on suggesting communities that users are interested in. In the process of developing the model, we propose a methodology for building semantic-based recommendation systems. In terms of evaluation, we performed two experiments using a developed mobile application called Dino. First, we presented hypothetical scenarios based on possible real-world SSN applications to measure users’ perceived sense of community. Second, we asked users to consider their real interests in suggested groups. We computed average values of 0.72 and 0.83 for precision and recall, respectively. The experiments’ results to assess the proposed scenarios ascertain average values of agreement above 80% for sense of community, belonging, social usefulness, member loyalty, and communities’ ephemerality. Therefore, our evaluation depicts that dynamic virtual communities formed by a SSN model based application would beneficially improve a social-aware virtual environment.

Expert Systems with Application