A Text Sentimental Approach for Online Portals Using Hadoop

Big data is an emerging technology to process the vast amount of both structured and unstructured data. Now a day social media such as twitter, face book, blogs and forums are the well suitable source to gathering the huge amount of data. Text sentiment analysis for the online portals such as flip kart, Amazon, Godaddy, etc.. are very important to review about their product performance in the market. Sentiment analysis is a text analysis method which aims to contextualize the meaning of the social network data. In the existing work, sentiment analysis is done by polarizing the sentences which derived from the public opinion. However it cannot polarize the public opinion accurately where the sentiment analysis is performed over the social network data"s. In this work, we target on finding an appropriate polarity recognition method for public opinion supervision system. In our method, we explore new feature extraction rules which extract emotional nouns, verbs, adjectives, and bigrams as representative features. Then, we apply Fuzzy Naive Bias to classify these online opinions into positive and negative class. Also we introduce the new category of sentiment analysis namely called as "Neutral". The experimental conducted were proves that the proposed methodology provides better result than the existing methodology