Video storage on the basis of content is the problem under consideration. The research addresses classifying video
content into clusters and how this classification can be used for business intelligence and other content guided parameterization. With fast growing concepts of video share and uploads, it has become difficult to store and retrieve
videos in an effective manner and how to perform filtering for videos that do not fit into the classified search item or in that particular cluster.
The problem does not end at creating clusters, it magnifies on how many clusters can be formed, what type of clusters will there be, how can a cluster own a piece of information and not manipulate it. How well a system can identify type and number of cluster, the following example highlights the mentioned problem
Lets suppose for time scale of 10 seconds starting at T=0; till T=9, there are 100 video uploads which have no prior information associated with them. No tag or context sensitive information has been provided with it. The concern is how the algorithm will create clusters and store according to information rather than associated words, texts or tags.