Stylometric Analysis for Authorship Attribution on Twitter

Authorship Attribution (AA), the science of inferring an author for a given piece of text based on its characteristics is a problem with a long history. In this paper, we study the problem of authorship attribution for forensic purposes and present machine learning techniques and stylometric features of the authors that enable authorship to be determined at rates significantly better than chance for texts of 140 characters or less. This analysis targets the micro-blogging site Twitter, where people share their interests and thoughts in form of short messages called "tweets". Millions of "tweets" are posted daily via this service and the possibility of sharing sensitive and illegitimate text cannot be ruled out. The technique discussed in this paper is a two stage process, where in the first stage, stylometric information is extracted from the collected dataset and in the second stage different classification algorithms are trained to predict authors of unseen text. The effort is towards maximizing the accuracy of predictions with optimum amount of data and users under consideration.