Florida International University
Ramesh Baral is a Ph.D. candidate in School of Computing and Information Sciences at FIU under the supervision of Professor Sitharama S. Iyengar. His preliminary research was done under the supervision of Professor Tao Li. Ramesh received his B.E. in Software Engineering from Pokhara University, Nepal in 2007. Before joining FIU in 2013, he was a Senior Software Engineer at Verisk Health Inc. Nepal for five years. His research interests include social network analysis, data mining, and machine learning. His research contributions have been published in renowned conferences and journals, such as ACM RecSys, IEE IRI, and Journal of Data Mining and Knowledge Discovery.
The increasing volume of information has created overwhelming challenges to extract the relevant items manually. Fortunately, the online systems, such as e-commerce (e.g., Amazon), location-based social networks (LBSNs) (e.g., Facebook) have the ability to track end users’ browsing and consumption experiences. Such explicit experiences (e.g., ratings) and many implicit contexts (e.g., social, spatial, temporal, and categorical) are useful in preference elicitation and recommendation. As an emerging branch of information retrieval, the recommendation systems are already popular in many domains, such as movies (e.g., YouTube), music (e.g., Pandora), and Point-of-Interest (POI) (e.g., Yelp).
The POI domain has many contextual challenges (e.g., spatial (preferences to a near place), social(e.g., friend’s influence), temporal (e.g., popularity at certain time), categorical (similar preferences to places with same category), locality of POI, etc.) that can be crucial for an efficient recommendation. A detailed analysis of the impact of such major contexts and their incorporation for POI recommendation is still a viable research direction. The check-in trends of users and influences of POIs on a region are also crucial for an efficient recommendation. Such influences and activity trends can be mapped into a latent space and contextually extended for preference elicitation.
The variation of contextual preferences implies different preference trends in different regions. Another viable problem is to efficiently aggregate such preference trends to model the locality preferences of users. Though many recommenders exist, most of them are less transparent and non-interpretable (as they conceal the reason behind recommendation). We can exploit the users’ reviews to extract aspects and generate explainable recommendations. In this study, we focus on the major contexts of LBSNs and present some novel techniques to model multi-context, locality-aware, and explainable POI recommendation.