Context-aware Youtube recommender system

摘要

Youtube is one of the most popular video sharing online resource that has millions of users around the world. The huge bulk of videos, which are growing at a high rate is posing problems for users to traverse through to relevant content. Users are facilitated with recommended videos that appeal to there interests. Following a hybrid recommendation approach, videos are recommended based on both collaborative recommendation and content-based recommendation. A limitation associated to this approach is that the videos recommended may not necessarily be appropriate to the current context that the user is in. Its very common for a single user to follow different interests depending of on the context they are in. A context-aware recommender system is proposed for Youtube that keeps track of multiple interests of a user and recommends videos based on their current context only. It serves a user better in finding relevant videos and has higher relevance to human judgment.