Sentiment Analysis Approaches

Lexicon-based methods require large labeled training datasets Lexicon-based methods cannot handle context-dependent sentiment, sarcasm, or domain-specific language well Lexicon-based methods are too slow for real-time use Lexicon-based methods can only classify into positive and negative, never neutral