An Extensive Study on Masked POS Tagging for the Telugu Language

Part-of-speech (POS) tagging is a crucial task in natural language processing, particularly for low-resource languages like Telugu. This study explores POS tag prediction in Telugu using masking techniques. This work assesses both traditional machine learning models—Random Forest, Decision Trees, and XGBoost—as well as deep learning models, including RNN and LSTM. The models are assessed based on accuracy, precision, recall, and F1-score for each tag. The results indicate deep learning models, particularly LSTM, outperform traditional machine learning approaches in capturing Telugu language patterns and improving POS tagging accuracy. This study addresses key important issues in POS tagging for low-resource languages and offers valuable insights to enhance natural language processing for Telugu.