Sentiment Analysis
An NLP task that determines the emotional tone or opinion expressed in text, classifying it as positive, negative, neutral, or along more nuanced dimensions.
Approaches
Fine-tuned BERT/RoBERTa classifiers, LLM prompting (ask the model to classify sentiment), lexicon-based methods (word lists with sentiment scores), and aspect-based analysis (sentiment toward specific features).
Applications
Brand monitoring, product review analysis, social media analytics, customer feedback analysis, financial news sentiment (for trading signals), and political opinion tracking.