Sentiment Analysis

Sentiment analysis — also called opinion mining — is the task of determining the emotional tone expressed in a piece of text. At the coarsest level this is a three-class classification: positive, negative, or neutral. More fine-grained approaches detect specific emotions (anger, joy, surprise), assign a continuous sentiment score, or perform aspect-based sentiment analysis to identify the sentiment toward particular attributes of a product or service.

Early systems used lexicon-based approaches: a sentiment dictionary assigns polarity scores to words, and the scores are aggregated across the document. These are fast and interpretable but fail on negation ('not good'), sarcasm, and domain-specific language. Machine learning classifiers trained on labeled reviews improved accuracy substantially, and transformer-based models now achieve near-human performance on standard benchmarks.

Sentiment analysis is widely applied in brand monitoring, customer feedback processing, financial news analysis, and social media tracking. Despite its maturity, the task remains difficult in practice: irony, mixed sentiment within a single document, and cultural variation in emotional expression continue to challenge even the best models.