Conceptual Analysis for Timely Social Media-Informed Personalized Recommendations

Integrating sensor networks and human social networks can provide rich data for many consumer applications. Conceptual analysis offers a way to reason about real-world concepts, which can assist in discovering hidden knowledge from the fused data. Knowledge discovered from such data can be used to provide mobile users with location-based, personalized and timely recommendations. Taking a multi-tier approach that separates concerns of data gathering, representation, aggregation and analysis, this paper presents a conceptual analysis framework that takes unified aggregated data as an input and generates semantically meaningful knowledge as an output. Preliminary experiments suggest that a fusion of sensor network and social media data improves the overall results compared to using either source of data alone.