Ammar Alanazi

[A Probabilistic Reciprocal Recommender with Temporal Dynamics] Recommender systems are methods of personalisation that provide online services users with suggestions for further interaction with those services. Most recommender systems are for product-to-consumer recommendation, suggesting items or products to users, but there is a growing need for reciprocal recommenders, where the goal is to suggest users to other users of the system with whom they might like to interact. Unlike product-to-consumer recommendation, for reciprocal recommendation to be successful both parties must agree to the interaction. Reciprocal recommendation is needed, for example, in matching users in online social networks, or matching users to employers in online recruitment. Since online social networks typically have many millions of users, the problem of finding successful interactions is not a trivial task. Most existing reciprocal recommenders assume that user preferences are static and ignore temporal aspects of user behaviour. In this research we take a different approach, and addresses the issue of representing user preferences as dynamic. We introduce a novel representation of observations in the dataset as a sequence of events instead of the traditional methods that consider observations as isolated events. Then we propose a Hidden Markov model (HMM) content-based reciprocal recommender that uses this new representation to generate recommendations. This HMM recommender was evaluated on the real-world dataset and results show that it improves recommendations compared to existing systems that do not use an explicit temporal representation. Finally, we develop a general framework for combining the HMM recommender with CF techniques in a unified hybrid recommender. Additionally, a new similarity measure is introduced to rank the hybrid recommender’s recommendations. Moreover, we implement a reciprocal recommender using the suggested framework and similarity measure. Evaluation of this system shows that it generates better recommendations than existing systems in a time-efficient manner.