Spontaneous social network: creating dynamic virtual communities based on context-aware computing

Abstract

With the emerging of online social networks along with the worldwide diffusion of smartphones, context awareness has become an essential concept in the field of mobile computing. Recent efforts and relevant research 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 novel Mobile Social Network (MSN) model called Spontaneous Social Network (SSN). 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 the literature, we found several works that lack on the community grouping approach, on the aspect that they are either limited to a specific location, or do not fully support virtual social interactions. We develop a mobile application called Dino, to provide a glimpse of what an SSN based application would be. To evaluate our model we perform two experiments using the developed mobile client. First, we present hypothetical scenarios based on possible real-world SSN applications to measure users’ perceived sense of community. The scenarios described are (1) music concert (2) sport event (3) shopping mall (4) conference or workshop (5) school or university. Second, we ask users to consider their real interests to assess our formed groups regarding their relevance and measure precision and recall of the groups’ suggestions. We compute 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 of 84% for sense of community, 80% for sense of belonging, 90% for social usefulness, 92% for member loyalty, and 81% for communities’ ephemerality. Therefore, our evaluation depict that dynamic virtual communities formed by a SSN model based application would beneficially improve a social-aware virtual environment.