Understanding video interactions in youtube

Abstract

This paper seeks understanding the user behavior in a social network created essentially by video interactions. We present a characterization of a social network created by the video interactions among users on YouTube, a popular social networking video sharing system. Our results uncover typical user behavioral patterns as well as show evidences of anti-social behavior such as selfpromotion and other types of content pollution.

Human Perception of Near-Duplicate Videos

“…From our results, human perception of NDVC matches many of the features present in its technical definitions with respect to manipulations of non-semantic features [2,4]. However, it is yet not clear whether similar clips differing in overlaid or added visual content with additional information can be considered as near-duplicates.…”

“…In the last few years, different research groups have tried to understand how videosharing web sites are used and in particular the impact that near-duplicate video clips (NDVC) have on video information retrieval tasks [4], spam creation [1] and identification of copyright infringements [2]. Most of the previous work has focused on identifying and removing NDVC.…”

“…According to Wu and colleagues [4], near duplicate videos differ in file format or encoding parameters, photometric variations (e.g., color) and editing operations (e.g., overlay, captions, add/remove scenes). Other scholars have employed an extended definition of similarity between features including changes on capturing time [2], for instance a different camera viewpoint. Finally, Basharat et al included similarity at the semantic level [1], where the same semantic concept can occur under different scene settings (e.g., two videos of different deer in different forests grazing moss).…”

Call Me Guru: User Categories and Large-Scale Behavior in YouTube

“…In contrast, very few works in computer science have been devoted to the study of user behavior on large-scale data samples gathered using YouTube's API or web crawlers. Using statistical methodologies, they have focused on analyzing the structure and topology of the social networks that emerge from user interaction, to reveal network characterstics that are relevant for information and communications services [18], and to detect anti-social and spam behaviors [3]. Despite the relevance of their findings, these works do not provide any understanding on the nature of users' behaviors and their motivations to participate and to create and maintain self and group identities.…”

“…However, few works have focused on characterizing long-term user behavior Existing works have provided a brief statistical analysis on the use of YouTube social-oriented features [11], studied the properties of the social network of friends and subscriptions [18], and a similar characterization based on user video interactions [3]. Understanding how users typically behave and interact, and how they perceive YouTube as both a system and a social outlet is fundamental to improve its performance.…”