Comparison of Individual Versus Stereotype Users in Tourist Recommender System using Agent Based Modeling

Abstract

Information overloading is a well-known issue across many domain. This problem is also pertinent in spatial domain. The issue can be resolved through recommendation and personalization techniques. Recommendation is an established approach in several domains including tourism. The proposed study is aimed for tourism in particular comparing individual versus stereotype user’s models through simulation of implicit as well as explicit user profiling by utilizing Spatial Agent Based Modeling. The Agent Based Modeling technique is employed with recommender system to cater information overloading problem in tourism. The work presented in this thesis presents a proof of concept where Spatial Agent based modeling can help in decision making. The results show that the recommendations done through individual user modeling are less accurate then dynamic stereotype user modeling. In Pakistan’s context work on agent based modeling as well as recommender system is remain unexplored. However this work can be a foundation step for other domains of application in Pakistan. This research area has its own importance due to its power of handling information overloading by using recommender as well as performing any type of mimic experiment using agent based modeling.