On the prediction of re-tweeting activities in social networks - A report on WISE 2012 challenge

Abstract

This paper reports on our participation in the Data Mining track of the WISE 2012 Challenge. The challenge is to predict the volume of future re-tweets and possible views for 33 given original short messages (tweets). Towards this, we compare and contrast four different methods and highlight our methods of choice for accomplishing this challenge. The first method is a naïve approach that discovers a regression function based on the popularity of messages and network connectivity. The second approach is to build a classifier that learns a classification model based on the user's preferences in different categories of topics. The third approach focuses on a network simulation that leverages a Monte Carlo method to simulate re-tweeting paths starting from a root message. The fourth approach uses collaborative filtering to build a recommendation model. The results of these four methods are compared in terms of their effectiveness and efficiency. Finally, insights into predicting message spreading in social networks are also given. © 2012 Springer-Verlag.