There are multiple multimedia devices and multiple users in a house. This paper presents research in media recommendation (movies, TV shows) for users in a home environment. The purpose is to produce the best recommendation based on the common interests of the users. The authors describe a research project called optimized multimedia service (OMUS) system.
The first problem that is considered in the paper is how to deal with the content that is available on various devices. The authors mention the Digital Living Network Alliance (DLNA), the organization established by Sony and Intel in June 2003 that defines standards to enable the sharing of digital media. DLNA is using the set of networking protocols defined by Universal Plug and Play (UPnP).
The section on recommendation outlines three categories: (1) collaborative filtering, which finds similar users; (2) context-based filtering, using the history of previous interests of the user; and (3) knowledge-based filtering, such as “this category is interesting for male users aged between 20 and 25.”
The next topic that is considered in this work is the support for group recommendations, based on user feedback or common preferences.
The OMUS information system architecture is presented in detail, with sections on the content aggregation, the user interface, and recommender systems. The OMUS system is compared with existing home information systems: YouTube, iTunes, and Kodi (formerly XBMC).
This 30-page paper deals with an interesting new domain. It has good scientific value and also shows a practical implementation. It could be a basis for advanced new products in this area.