Comparative Analysis of Sentiment Analysis Between All Bigrams and Selective Adverb/Adjective Bigrams

摘要

Increase in the amount of unstructured data across different platforms serves as a valuable resource for predicting market trends, analyzing product features, and considering the customer sentiment in designing new features/products. The sentiment of unstructured data such as tweets, Facebook comments, and web reviews is calculated by using the polarity and intensity of the words, whereas polarity indicates positive or negative sentiment, and intensity indicates the strength of polarity. In this paper, a comparative study of sentiment analysis performance and accuracy between all bigrams and selective adverb/adjective bigrams is done. The outcome of this research will serve as a metric for both academia and industry to implement sentiment analysis projects.