The Impact of Oath Writing Style on Stylometric Features and Machine Learning Classifiers

Abstract

Computational stylometry is the field that studies the distinctive style of a written text using computational tasks. The first task is how to define quantifiable measures in a text and the second is to classify the text into a predefined category. This study propose a stylometric features selection approach evaluated by machine learning algorithms to find the finest of the features and to study the impact of the features selection on the classifiers performance in the domain of oath statement in the Quranic text. The results show that better classifiers performance is highly affected by the best feature selection which is associated to an explicit oath style.