AsPOS: Assamese Part of Speech Tagger using Deep Learning Approach

TL;DR

This paper introduces a deep learning-based Part of Speech tagger for Assamese, achieving an F1 score of 86.52%, addressing resource scarcity in this low-resource language.

Contribution

It presents the first deep learning approach for Assamese POS tagging, evaluating multiple embeddings and establishing a baseline for future research.

Findings

Achieved 86.52% F1 score in POS tagging accuracy.

Evaluated multiple pre-trained word embeddings for Assamese.

Developed a two-phase training process for improved performance.

Abstract

Part of Speech (POS) tagging is crucial to Natural Language Processing (NLP). It is a well-studied topic in several resource-rich languages. However, the development of computational linguistic resources is still in its infancy despite the existence of numerous languages that are historically and literary rich. Assamese, an Indian scheduled language, spoken by more than 25 million people, falls under this category. In this paper, we present a Deep Learning (DL)-based POS tagger for Assamese. The development process is divided into two stages. In the first phase, several pre-trained word embeddings are employed to train several tagging models. This allows us to evaluate the performance of the word embeddings in the POS tagging task. The top-performing model from the first phase is employed to annotate another set of new sentences. In the second phase, the model is trained further using the fresh dataset. Finally, we attain a tagging accuracy of 86.52% in F1 score. The model may serve as a baseline for further study on DL-based Assamese POS tagging.