Sentiment Analysis with Word Embedding

The basic task of sentiment analysis is to determine the sentiment polarity (positivity, neutrality or negativity) of a piece text. The traditional bag-of-words models deficiencies affect the accuracy of sentiment classifications. The purpose of this study is to improve the accuracy of the sentiment classification by employing the concept of word embedding. This study uses Word2Vec to produce high-dimensional word vectors that learn contextual information of words. The resulting word vectors are used to train machine learning algorithms in the form of classifiers for sentiment classification. Our experiments on real world datasets shows that the use of word embedding improves the accuracy of sentiments classification.