Dynamic multi-concept user profile modelling in research paper recommender systems
- Modhi Al Alshaikh(Author)
- University of Brighton
Student thesis:
Student thesis
Doctoral Thesis
About the thesis
The internet and the digital libraries are major sources of information for<br/>researchers, and there is an enormous growth of information on these sources. A large<br/>number of research papers are available which leads to the information overload<br/>problem and hence finding research papers that are related to users’ interests become<br/>difficult and time consuming. The field of recommender systems aims to solve the<br/>information overload problem by filtering information and providing users with<br/>relevant results. Although the current recommender systems provide recommendation<br/>services to users, different limitations and challenges have not been adequately<br/>addressed in the research paper domain. The work presented in this thesis contributes<br/>to the development of models and algorithms to the recommender systems in the<br/>research paper domain. The main aim of this thesis is to develop a dynamic multiconcept<br/>system that is able to recommend research papers of interest at appropriate<br/>times. The first contribution of this thesis is modelling dynamic user profiles that are<br/>able to adapt to the changes in multiple user interests and to be compatible with the<br/>requirements of advanced ontologies. The second contribution is analysing users’<br/>reading behaviour with research papers to develop novel short-term and long-term<br/>models that are able to adapt dynamically according to a user’s changing behaviour<br/>during his/her short and long term goals. These models can effectively learn different<br/>users’ reading behaviours implicitly without the need for any intervention from the<br/>user. The third contribution is predicting user’s future interests using a novel<br/>collaborative filtering approach without the need for the user ratings. All our proposed<br/>models are evaluated using offline evaluations with the BibSonomy dataset that<br/>contains actual users’ records. Our results show that our models outperform the<br/>baselines used for comparisons. Finally, we integrated our models to one unified<br/>dynamic hybrid system in order to provide recommendations which most closely<br/>represent the users’ research interests at particular times. The evaluation results<br/>indicate that the dynamic hybrid system that models and integrates multiple user<br/>interests and concepts can bring substantial benefits to a recommender system in the<br/>research paper domain.
Thesis information
Thesis Award Date
2018Qualification Level
Doctoral ThesisOriginal Language
EnglishThesis Managed By
Supervisors
Gulden Uchyigit (Supervisor)
