Integration of morphological features and contextual weightage using monotonic chunk attention for part of speech tagging

Rajesh Kumar Mundotiya, Arpit Mehta, Rupjyoti Baruah, Anil Kumar Singh

Abstract: Highlights•The part of speech (POS) tagging problem aims to assign a grammatical category to the word that appeared in the sentence.•A monotonic chunk-wise attention is utilized to capture non-continuous long dependencies among the words of the sentence.•A monotonic chunk-wise attention with cnn-gru-softmax (MCCGS) architecture is presented using individual and handcrafted morphological features integration.•The morphological feature integration is performed at three distinct levels to compare the significance of feature learning for POS tagging.•This study concludes with extensive experiments to validate the performance of proposed architecture against the state-of-art- architectures.