SRMCS: A semantic-aware recommendation framework for mobile crowd sensing

With the rapidly growing number of wireless sensors and mobile phones, massive amounts of semantic-aware sensory data are being produced at every moment. Meanwhile, an overload problem has been generated in insular systems and has caused semantic-aware information to be ineffectively utilised. A central challenge for cognitive information processing is to design a scheme that can dispatch semantic-aware information to the appropriate context and to push it to users who are truly interested. This study proposes a framework for semantic-aware recommendations for mobile crowd sensing (SRMCS), which focuses on the relationships among sensory information rather than the information itself, and is based on a multi-layer graphic model (MLGM) that we propose in this study. We leverage semantic relatedness from different dimensions to construct contexts, allowing semantic-aware information to disperse into different relative contexts to obtain a better perception of the real world for the users. We implement the ability to push semantic-aware information for recommendations in terms of individual preferences, demands and social relationships, allowing the users to obtain more valuable semantic-aware information. The users can design different evaluation mechanisms with the objective of applying the framework to different occasions and to evaluate its effect. The framework that we propose is applied to one recommendation test study; the evaluation demonstrates the power of its reasonableness and effectiveness.