On the Usage of Neural POS Taggers for Shakespearean Literature in Social Systems
Avinash Samantra, Pankaj Kumar, Tu N. Nguyen, Arun Kumar Sangaiah, Sambit Bakshi · IEEE Transactions on Computational Social Systems · 2022
Part-of-speech (POS) taggers are the primary requisite of any natural language processing (NLP) mechanism. Conventional POS tagger and libraries are expert-made or static and concentrate on the literature domain. These POS taggers limit the performance of subsequent mechanisms like polarity detection, sentiment analysis, opinion mining, and so on. The unsuitability of a tagger for a new genre of literature makes famous libraries, such as Natural Language Toolkit (NLTK) and University Centre for Computer Corpus Research on Language (UCREL) Constituent Likelihood Automatic Word-tagging System Seven (CLAWS7) create the need for a neural POS tagger to serve Shakespearean literature. This article reports a preliminary study on the suitability of the neural taggers over static or manual taggers, supported by the accuracy of 97% achieved onHamlet. Furthermore, these neural networks are scalable over the literature domains irrespective of the stylistic variations, opening up this area to computer scientists to aid literary enthusiasts in contributing to domain of the social systems.