An AI-driven social media recommender system leveraging smartphone and IoT data

Our research presents “RoBERTaRecIOT,” an innovative model that stands out for its superiority. It utilizes the pre-trained Robustly Optimized Bidirectional Encoder Representations from Transformers (RoBERTa) framework to deliver personalized travel recommendations via social media. This model is not just a theoretical concept but a practical solution, proposing an advanced travel route recommendation system that automatically gathers tourists’ onsite behavioral data related to specific POIs, utilizing a smartphone and the Internet of Things (IoT) technologies. It surpasses traditional sequential prediction barriers by integrating bidirectional context, non-symmetric schemas, and sophisticated enhancing user similarity evaluations through topic modeling. Furthermore, we introduce a new preference assessment method employing explicit demographic information, significantly mitigating the effects of the cold start issue. Our empirical studies, utilizing Yelp and Flickr datasets, demonstrate the model’s superiority, surpassing conventional metrics with improved F-Score, MAP, and NDCG values. The “RoBERTaRecIOT” model stands out as a revolutionary instrument in the domain of social media-driven travel recommendations, providing a dynamic and user-focused experience for global adventurers.