Sistema de recomendação personalizada para ambientes de TV digital

Abstract

The advent of digital TV brought with it new horizons for the TV watchers, full of new opportunities and possibilities. They will have at their disposal a wide variety of information. On the other hand, a new problem arises: information overload. The present dissertation presents a approach for the Digital TV environment, aiming at lowering information overload. By using customing and modeling techniques, it is shown a dynamic way of relating information shown on TV and the watcher‘s profile. Some advise techniques were used in order to offer a customized scheduling to the watcher, aiming at satisfying his interests. Stereotypes proved necessary as well, in oder to compensate the unavailability of data about the watcher, in order to avoid cold-start problem. This dissertation therefore has as its key objective the proposal of a custom scheduling to the Digital TV watcher.