Next Point of Interest Recommendation Using Adaptive Weights for Specific Behavioral Patterns

Next Point of Interest (POI) recommendation plays a pivotal role in assisting users in discovering relevant and enjoyable destinations during their travels by analyzing historical check-in data. However, a significant challenge arises when users are travelling unfamiliar territories where their preferences and contextual factors may differ from daily life pattern. The pattern of user behavior when traveling to new places is dynamic. Furthermore, we cannot import data from local pattern traveling to compensate for this scenario. To address this issue, we propose the Next Point of Interest Recommendation Using Adaptive Weights for Specific Behavioral Patterns (AWSBP) to learn user preferences for two patterns of users. Our focus is on rearranging and structuring the dataset to optimize model performance and the relevance of recommendations. Rigorous measures are implemented in this research to uphold user privacy; only anonymized location data is utilized, ensuring that no user profile information is accessible or analyzed. This safeguards the confidentiality and anonymity of individuals’ personal data from two real-world datasets. The AWSBP framework outperforms state-of-the-art models in terms of Recall and NDCG metrics for next POI recommendation.