Publications
Also on Google Scholar.
COVID-19-related Nepali Tweets Classification in a Low Resource Setting
Proceedings of The Seventh Workshop on Social Media Mining for Health Applications, Workshop & Shared Task · 2022
Identified eight COVID-19 discussion topics in Nepali-language Twitter. Built an automated pipeline to gather,
classify, and visualize tweets in Nepali and Devanagari scripts. Compared mBERT and MuRIL for low-resource
tweet classification — results showed MuRIL outperforms mBERT at larger data sizes. Data, models,
and dashboard are open-sourced.