Individualizing user profile from viewing logs of several people for TV program recommendation
Koichi Iguchi, Yoshinori Hijikata, Shogo Nishida · 2015
Recommendation services for e-commerce and content distribution sites are increasing in popularity. An individual is identified by a login certification, and then information pertaining to their interests and preferences are collected. Finally, the recommendation service recommends items to the user based on the information gathered. Recommendation services are not as popular for recommending television programs as they are for Web-based services. The primary reason for this discrepancy is that it is difficult to identify an individual sitting in front of a television, which is usually shared by several members of a family. Viewers generally do not want to perform additional operations on the remote control for login authentication rendering it impossible to create a user profile that contains personal interests and preferences. Therefore, in this study, we propose a technique to individualize a user profile from the combined viewing logs of several television viewers.