Context-Aware Recommendation System Survey: Recommendation When Adding Contextual Information

The recommendation system (RecSys) helps users provide a referral by filtering information based on the relationship between interests and needs. Several methods and approaches in the SR domain, such as CBF, CF, and hybrid, continue to be developed in looking for recommendations relevant to user interests. Recent research suggests that adding contextual information can significantly increase the accuracy of RecSys. However, adding contextual information to RecSys also raises new problems, such as increasingly sparse data. The context-aware recommender system (CARS) is a solution to overcome the problem of changing user preferences by entering contextual information into the recommendation method. In this survey paper, we try to define and classify the types of contextual information. Contextual attributes appropriate for the domain under discussion can be captured using our definition and classification of contextual information. In addition, we also describe and summarize the modifications made by researchers to the model. We also describe the evaluation method and dataset that can be used in the domain of CARS. Finally, we present future research in three schemes at the end of this survey paper.