Abstract
Conference Title: 2016 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM) Conference Start Date: 2016, Aug. 18 Conference End Date: 2016, Aug. 21 Conference Location: San Francisco, CA, USA With the proliferation of social networks and their popularity especially Facebook and Twitter, healthcare related issues arise. Many users seek to follow and discuss similar disease experience. Thus, it sounds promising to develop a friendship prediction system for healthcare purposes. Moreover, due to the dynamic aspect of social networks link (friendship) prediction has attracted the attention of many researchers. Link prediction consists on inferring probable links that may occur in the next time-stamp. Shortcomings of this task reside on that links are predicted according to only one time period and advanced privacy settings are behind the invisibility of interactions. In this work, we present an overview of some existing link prediction methods. We mainly propose a new approach to predict possible friendship between individuals and also we integrate fuzzy logic to deal with the lack of precision and the vagueness in the similarity between two individuals. Finally, in order to validate our solution, we use real data. The experiments show encouraging results. Comparing with crisp approaches, fuzzy method seems more effective and shows more accurate predictions.