Automatic Categorized Corpus Creation of Hindi Poetries Based on Rasa(s) for Linguistics Research

Corpus creation is required for solving natural language processing problems and linguistic research using machine learning and deep learning. Classification problems solved using supervised learning methods need labeled corpus. It is tedious task if done manually. This research article is creating labeled corpus of Hindi poetries based on Rasa(s). Preprocessing of the extracted documents is done using NLP techniques to create a clean corpus ready to use for solving classification problem or any other problem. The garbage elements are removed from the extracted files, and file containing excess garbage is deleted. The corpus created comprises of poetries of 9 Rasa(s) also called Navrasa. A total of 1089 files were extracted comprising of 333 Adbhuta, 5 Bhayank, 25 Hasya, 70 Karuna, 6 Raudra, 277 Shanta, 218 Shringar, 147 Veera, and 8 Vibhasta Rasa poetries in 04:20:23 hrs with negligible human intervention. A total of 17 garbage files were deleted, and the final corpus have 1072 poetry documents comprising of 328 Adbhuta, 5 Bhayank, 25 Hasya, 68 Karuna, 6 Raudra, 272 Shanta, 216 Shringar, 144 Veera, and 8 Vibhasta Rasa poetries.