A weighted semantic feature expansion using hyponymy tree for feature integration in sentiment analysis

摘要

It is necessary that sophisticated opinion extraction methods to be used, since ordinary keyword search does not suit for mining varied opinions. Assessment, proposed emotional correspondence or compelling state are might be his or her assessment is attitude. Classifying polarity of an unstructured document text in terms of positive, negative, or neutral polarities is the task of sentiment analysis. This paper proposes semantic based feature integration for feature extraction. Experiments were undertaken with different features.