MODELING ACTIVE LIFE SPAN OF YOUTUBE VIDEOS BASED ON CHANGING VIEWERSHIP- RATE

Abstract

Among the various social media sites, YouTube has emerged as a forerunner in the race of online video sharing sites. Being a free uploading and viewing site, YouTube’s and the video uploader’s main source of revenue are the advertisements run before and during the video. Thus, to better optimize the profits, modeling the view-counts of a particular video is very essential. It is a known fact that the view-count on a video increases with time but the view-count growth rate (viewership rate) is not always
increasing. It increases rapidly during the time period when the video becomes viral and it again slows down once this virality
phase is over. The behaviour of viewership rate changes multiple times throughout the video life-cycle and leads to altered
lifespan of the videos. In the current proposal we have predicted the number of view counts on the basis of changing viewership
rate. This change however, might occur several times during video lifecycle, therefore, this ideology has been incorporated in
proposed methodical work. A set of models have been discussed which have been ranked using “VIKOR”- multi criterion decision making (MCDM) technique using the YouTube video data sets.