A Video Sharing Site Oriented Data Collection System

Abstract

Analysis of video data have significant meaning for studying the characteristics video sharing site and understanding its user's behavior and preference. In order to achieve this goal, we designed and implemented a high performance video sharing site oriented data collection system (VDCS), which overcomes the low efficiency problems that traditional crawler facing. High time efficiency video search strategy is proposed to extract structured video data and high space efficiency filter mechanism is explored for filtering videos. The experiments and analysis proves VDCS have much more practical significance in collecting video data compare with traditional crawler due to its video search strategy and filter mechanism. Through analysis of the datasets we built, we studied video sharing sites' characteristic which includes the upload trend of videos, views and comments and social networks among videos, this study enhanced our understanding of video sharing sites and provide a new measurement tool for analyzing of video data.