Automated Generation and Configuration of Domain-Specific Recommender Systems

Recommender systems (RSs) are tools that aim to assist end-users in various sectors, spanning from tourism to e-commerce. However, the development process of such systems is challenging, since they should suit different domains and rely on diverse data. This brings inefficiencies, thus more development time, also due to the lack of configuration approaches. This work presents the ongoing development of a PhD thesis about a methodology for leveraging Model-Driven Engineering (MDE) and model-weaving techniques to facilitate the development of easily configurable RSs. This approach is intended for reuse enhancement, error reduction, and time-to-deploy reduction. The current achieved results underline the gap in the literature and the lack of configurable support systems that RSs providers need, pointing out the need for future research to develop methodologies that can treat dynamic and context-aware data effectively.