Built a cross-lingual sentiment analysis model over 8+ languages for monolingual and cross-lingual tasks, achieving Sentiment Graph F1 of 0.55 (cross-lingual) and 0.559 (monolingual) using RoBERTa with CRF for span extraction and polarity classification.
python nlp machine-learning ml nlp-machine-learning sentiment-classification nlp-projects structured-sentiment-analysis iiitd-nlp
-
Updated
Feb 25, 2026 - Python