With hundreds of TV channels, a good TV program recommendation system can save time. Hadoop fair scheduler cloud computing is designed to make information processing and filtering effective and scalable. In cloud computing, computers are connected over a network and perform computation simultaneously; more computation power can be obtained by adding more computer nodes. In the present study, cloud computing is used to build a TV program recommendation system. A fair scheduler cloud structure is applied to improve the system performance. For program recommendation, the K-means recursive clustering algorithm is used for user clustering, the term frequency/inverse document frequency algorithm is applied for finding related popular programs, and k-nearest neighbor is used to recommend programs. Most TV program recommendation systems focus on providing a personal recommendation system. The proposed system also considers user groups and the program watching preferences of the majority. The proposed fair scheduler cloud-based architecture is scalable; a massive amount of information can be processed in real-time to obtain program recommendation results that can represent almost all users.
All Science Journal Classification (ASJC) codes