Social networking has become a major part in internet. Speed of data shared is in the form of text, images and videos. The Online videos are shared mostly in popular websites such as Youtube, facebook, etc and quiet a lot of time is wasted by the users in acquiring to search their required media. Usually recommendations are provided based on history and user behavior, while results of searches given are not accurate and generally produces huge computations complexity. Therefore using a cloud-based collaborative multimedia recommendation system which can reduce network overhead and speed up the recommendation process. The user?s upload the video with specific keywords and later collaborative filtering is applied. Social websites have become a growing influx of unrelated contents. Video search systems can be fooled by malicious attacks. Thus spam removal methods are used in this work. The detection of spammers is done using customer suggestion private storage formation using classified algorithms. This approach succeeds at separating spammers and promoters videos. The delay is reduced as search is obtained easily and spammers are removed to an extent.