Mining quad closure patterns in Instagram

Abstract

Quad Closure is a group of four people who are connected with each other. In this paper, we propose a new group recognition method for Instagram which are based on triadic closure method to determine groups on dynamic social networks (e.g likes and comments) between users as named Quad Closure. Social networks are not easily classified because of their complexity. We study how an open quad closure becomes close quad and the possibility of finding this process. There are a wide variety of factors having effects on this process. These factors include users' background information like active using time piece, photos tag frequency and sentimental analyze on comments. We try to create a unique model to predict formation of quad closure with all these factors and we share our experimental results on data taken from Instagram and show the success of our method on quad closure patterns.