Domain-Specific Sentiment Classification

摘要: A domain-specific sentiment classifier that can be used to score the polarity and magnitude of expressed by documents is created. domain-independent lexicon established a uses documents. Sets high-sentiment having positive negative polarities are identified. The n-grams within filtered remove extremely common n-grams. saved as features in model. model trained using set training which may manually or automatically labeled their overall produce scores for lexicon. This classifier.