A comparison of classifiers and features for authorship authentication of social networking messages

摘要

This paper develops algorithms and investigates various classifiers to determine the authenticity of shortrnsocial network postings, an average of 20.6 words, from Facebook. This paper presents and discussesrnseveral experiments using a variety of classifiers. The goal of this research is to determine the degree tornwhich such postings can be authenticated as coming from the purported user and not from an intruder.rnVarious sets of stylometry and ad hoc social networking features were developed to categorize 9259 postsrnfrom 30 Facebook authors as authentic or non-authentic. An algorithm to utilize machine-learning classifiersrnfor investigating this problem is described, and an additional voting algorithm that combines three classifiersrnis investigated. This research is one of the first works that focused on authorship authentication in shortrnmessages, such as postings on social network sites. The challenges of applying traditional stylometryrntechniques on short messages are discussed. Experimental results demonstrate an average accuracy rate ofrn79.6% among 30 users. Further empirical analyses evaluate the effect of sample size, feature selection, userrnwriting style, and classification method on authorship authentication, indicating varying degrees of successrncompared with previous studies.