Sentiment Analysis on Huge Streaming Data using Hadoop

Abstract: Sentiment analysis has been a major focus in all consumer oriented industries due to the availability of huge amount of customer opinions in the Internet. This paper presents a sentiment analysis framework that utilizes the processing efficiency of the Hadoop ecosystem to provide real time sentiment analysis. The framework divides the process of sentiment analysis to two major sections; content pre-processing and evaluation. Experiments shows that our application has the ability to scale and handle huge amounts of data.

Keywords: Sentiment Analysis; Polarity Identification; Hadoop; Tokenization; Stemming