A Review on Prediction of Collective Behavior

Abstract—The social media provide platform to connect peoples with each other and they can share their views, idea, thoughts. Now a day lot of data are generated by social media. The heterogeneous peoples are like to connect with each other through the social sites such as Facebook, Twitter, Flickr, and YouTube etc, to classify this large data is also a challenging task. This paper is to study of predict collective behavior of individuals on the social network. For classification of large network into specific communities by consider pattern of social behaviour. The k-means variant edge-centric clustering algorithm is used for finding the communities of people. To classify the large network we are going to use the edge-centric clustering technique to extract the sparse social dimensions from network. The social dimension based approach has been shows beneficial for the study. With this approach we can handle millions of user data to demonstrate behavioural approach of user in social sites for marketing purpose