Recommender systems play a central role to reduce the problem of information overload and they propose techniques that are able to suggest products that are potentially interesting for certain users and groups. In the last few years, a new research topic has emerged with the goal of personalizing recommender results through the deep analysis of the social networks user links and diary interactions.

The main research goal of this project is to provide recommendation techniques based on advanced Artificial Intelligence techniques that allow the personalization of recommendation results, by exploiting the possibilities of using social knowledge connections. The proposal will focus on social aware recommendations, which show the value of using social networks and virtual communities as a means to increase the quality of recommendation processes for individuals and groups. More specifically, we explore the applicability of Case Based Reasoning, planning, Social Network Analysis, Data mining and sentiment analysis to acquire, formalize and use dinamically acquired knowledge that is integrated in the recommendation processes.

Our research will improve existing recommendation techniques by integrating social knowledge from the users. This social knowledge is captured by analysing how the users interact through different types of social networks: from close social circles (Whatsapp, Google+ Circles, Instagram) to wide scope social networks (like Facebook o Twitter) and using also opinion communities like TripAdvisor.

The use of mobile technology completes the capability of dynamically acquire additional contextual knowledge that allows the integration of the online experience with the real world (smart cities, hotels, monuments, cinemas, shopping). The advanced recommendation techniques will improve precision of the traditional recommenders, both for individual and groups and will transfer a direct benefit for the companies that integrate them in their business processes.

As a result of the project, we will include these advanced techniques into a platform for generating recommendation systems that offer society simple, efficient, and cheap solutions. This technology platform supports an agile methodology based on rapid prototyping and the computational evaluation of recommender systems by finding a configuration that behaves depending on the domain, the new algorithm and contextual information sources used. In addition, the platform will include support for the implementation and deployment of the prototypes and recommendation applications developed as web services, and client- and domain- independent platform, so that it will be made available to business recommendation capabilities with customization based on the intensive use of the social knowledge.

We facilitate the research transfer of the results of this project and validate the positive impact and enrich the contributions by affording development and experimenting in different real domains of application with real users. Recommender techniques are proven to be multidisciplinar and reusable through different prototipes related with recommenders in ecommerce, turism, leisure and teaching