Group Recommender System Using Hybrid Filtering for Tourism Domain

Abstract

Group recommender system is built to overcome the different needs that exist in each of the tour group members in determining tourist destinations. This is because everyone has different desires in determining the constraints of tourist destinations, such as costs and tourist categories. This recommendation process is based on a combination of profiles of each member to produce a tourist destination that can fulfill their needs. The method used is hybrid filtering which is a combination of collaborative filtering and knowledge-based filtering. Because the two methods have their own shortcomings. The collaborative method has a deficiency of tourist items that have never been rated by a user before, so the item cannot be recommended for active users. While in knowledge-based, tourist items must be completely described in detail so that they can be recommended to users. Domain used in this research is Bandung Raya area. The results of the evaluation in this study, users tend to feel very satisfied for the results of group travel recommendations generated the application by 50%. As for the evaluation of the recommendation algorithm, users prefer the results of the recommendations produced by collaborative and hybrid methods.