Computational Analysis of Rabindranath Tagore Poetry for Classification and Author Ship Attribution

Abstract

 Computational Analysis of poetry is always attracting many NLP scholars due to itschallenges.For the experts assessing the aesthetics’,style of poetry and poet is an expensive and cumbersome job.  In this paper we experimented with text data to extract the stylistic and linguistics information for analyzing, grouping and classifying the poetry. For this task we used semantic features to perform subject oriented clustering and classification of a renowned poet Rabindranath Tagore. We have removed the punctuations, numbers, stop words to clean the corpus and used TF-IDF weightage to find the frequent words for analysis purpose. The classification results are satisfactory. This kind of classification analysis is first attempt on Rabindranath Tagore poetry which is written in English.We extended this work by combining stylistic features with word features and performances of these two methods are analyzed. We compared his Epic poetry Geethanjali with other poems written by him.