Tourism Recommender System Based on Cognitive Similarity Between Cross-Cultural Users

Abstract

Nowadays, the speedy increasing information in tourism services since a massive amount of data is constructed by tourists experiences. The recommendation systems are widely applied to tourism services and focus on determining personalized user preferences to handle this extensive information. Exploiting the different cultural effects rarely consider in recent studies despite this factor influences recommendation based on user preferences. Furthermore, existing research only evaluates the relevance of cultural differences to their recommendation, rather than using the cross-cultural factors to recommendations systems. This paper proposes the collaborative filtering recommendation system based on similar tourist places where users from different cross-cultural can share their spatial experiences. To do that, we first collect user feedback about similar tourist places from many nationalities (consider as the cultures). We then exploit this feedback to define similar cross-cultural users (neighbors) based on a cognitive similarity. Finally, the system generates personalized recommendations based on user experiences and their neighbors. The initial dataset collected from TripAdvisor, consisting of four types such as hotels, restaurants, shopping malls, and attractions, is provided to the feedback collection function in our experiment. We were using the classical method, user-based Pearson correlation, as a baseline to demonstrate the performance of our proposed method. The result shows that the proposed system outperforms the baseline in terms of MAE and RMSE metrics.