A Recommendation System as a Digital Marketing tool for Online Communities

Recommender systems are able to predict users’ preferences and items of interest, by analysing historical data on their behaviour and actions. Different techniques exist and are applicable in different scenarios. This thesis explores how to combine Content-Based and Collaborative-Filtering techniques in a hybrid system and how personalised recommendations and one-to-one marketing techniques can lead to an improvement in user engagement. Specifically, it is analysed the case of online platforms where there is no rating system in place. Results are empirically tested and evaluated with training/testing approach and recommendations seem to be quite accurate. However, further online evaluation is needed to measure any actual increase in user engagement