This project focuses on document-level sentiment analysis of Norwegian texts, comparing three different models. LSTM, XGBoost and three different sizes of the Norwegian BERT model, NorBERT3.
The models are trained on the Norwegian Review Corpus, NoReC, which has been made specifically for the purpose of training and evaluating models for document-level sentiment analysis.
