Mobile Video Recommendation System on Cloud with User Behavior

Now a day, countless web videos are available on-line, but the problem is how to help users to find videos of their interest in an efficient way? Mobile user always wastes a lot of time to obtain their desired videos. The proposed system is a cloud-based mobile video recommendation system which can speed up the recommendation process and reduce network overhead. The Mobile properties are collected for context aware recommendation from video-sharing mobile application. The user based recommender system is created with the Mahout Machine learning library. Moreover, Mahout's core algorithms are used for classification, clustering and collaborative filtering. The System collects user, item, rating for creating clusters of user contexts and user profiles to generate video recommendation rules. User context and the profile clusters are used for finding video recommendation from the massive amount of video collection of cloud. Mahout provides better recommendation functions. Depending upon mobile properties and viewing time of video we create implicit preferences to generate recommendation rules. The system will provide a very good result with more accurate Recommendation Evaluation functions. The proposed system can recommend desired services with high recall, high precision and low response delay.