Sentiment classifiers based on feature extraction

摘要: Method and apparatus are provided for providing one or more sentiment classifiers from training data using supervised classification techniques based on features extracted the data. Training includes a plurality of units such as, but not limited to, documents, paragraphs, sentences, clauses. A feature extraction component extracts data, value determination determines each frequency at which occurs in On other hand, class labeling labels unit according to classes provide labeled Thereafter, classifier generation provides least technique.