Automatic authorship detection from Bengali text using stylometric approach
Nazmul Islam, Mohammed Moshiul Hoque, Mohammad Rajib Hossain · 2017
Authorship detection is the process of predicting authorship of an unknown text. Every writer has a different style of writing of their own. Detecting authorship from text by analyzing writing style of an author is known as stylometry. In this paper, we propose a stylometric feature based approach for detecting authorship from Bengali texts. The system the classify authorship using n-grams, a feature ranking and selection system using information gain (IG). We used 3125 passages written by 10 Bengali authors for evaluating performance. The evaluation result shows that the propose system achieved 96% accuracy in authorship detection using random forest classifier and also reveal that n-gram features are very good discriminators among linguistic style of different Bengali authors.