Integrating recommender information in social ecosystems decisions

The exploration of online social ecosystems whose members share mutual recommendations and interactions is a time-dependent and contextual-based process which aims to predict the social status among them. To address the difficulties associated with the process, this article presents the integration of the predictive recommender, social networks, and interaction components into a single methodology. The originality of the proposed framework stems from developing each model based on: (1) a time history and decay algorithm to consider temporal recommendations and interactions; (2) a predictive-aggregating function for different types of social contexts; and, (3) a homophily algorithm to evaluate people's interconnections proximity. Details of the framework are described, a recommender search strategy methodology integrating all of the above is devised, and a case study is used to demonstrate its capabilities. Possible extensions are then outlined.