The number of receivable TV channels has highly increased in recent years. Handling the enormous offer of TV content could be a challenge for the users in case of selecting the most interesting program. Furthermore, most users only focus on their favorite TV channels. That's the reason why content on other channels won't be recognized. For this problem assistive systems and tools are desirable to support in selecting the most appealing content with respect to the user's interests. The research group Next Generation PVR (NG-PVR) extends a personal video recorder (PVR) with a generic recommendation system based on a Bayesian classifier and adapted it for the use in the application area of television. The system analyzes the user's TV watching behavior to present new choices of content. So the system is able to generate personalized TV program recommendations. The content is stored on an internal hard disc drive where it is recorded for the user to watch. This paper presents the current state of development by introducing the system's architecture and implemented recommendation mechanisms.

No abstract