User profiling for content personalisation in information retrieval

One of the key issues with the overabundance of online information sources is that of finding what is relevant. The key to success for any type of information provider must be the personalisation of content in information retrieval, and this can be achieved through the maintenance of user profiles and the matching of these profiles to content metadata. This paper is concerned with user profiling and its role in content personalisation of information retrieval, and in particular presents a profile model which incorporates user preference information and action history information (representing the user’s previous searches). The benefits and costs of such a model are examined and it is argued that the benefits (including personalisation accuracy, computational costs extensibility and flexibility) far outweigh the costs. The matching of profiles to metadata is also discussed as it fulfils an important role in the personalisation process. Although, the user profile model presented is focused on E-Learning, the general platform could be applied to other areas.