In this paper, we study the problem of authorship identification in Bengali literary works. We considered three authors namely Rabindranath Tagore, Bankim Chandra Chattopadhyay and Sukanta Bhattacharyay. It was observed that simple unigram and bi-gram features along with vocabulary richness were rich enough to discriminate amongst these authors. Although results degraded slightly when training set size was considerably small. For larger training set, a classification accuracy of above 90% for unigram feature and almost 100% for bi-gram feature was achieved. Results could be improved further by using more sophisticated features. © 2011 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Das, S., & Mitra, P. (2011). Author identification in bengali literary works. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6744 LNCS, pp. 220–226). https://doi.org/10.1007/978-3-642-21786-9_37
Mendeley helps you to discover research relevant for your work.