Dynamic Packaging using a Cluster-based Demographic Filtering Approach

Dynamic packaging and product bundling are key topics in current tourism research and also heavily discussed within the travel industry. This paper describes a new approach to how a user model and a combination of collaborative and demographic filtering can be used to recommend product bundles in dynamic packaging. The user model differentiates between a short term component and a long term component. The short term component contains information about the user’s current session while the long term component saves data which holds beyond a session. Furthermore it is shown how stored data is used to recommend a combination of tourist services. This not only takes evaluations of other users into account, but uses demographic properties of the user as well.