A Comparative Study Of Sentiment Analysis For Big Data On Hadoop

Nowadays, people express almost all their feelings and views about everything in the surrounding world on social media applications. If these posts are analyzed accurately, they will be a huge opportunity for various organizations to increase their market value by using that information in decision-making. Opinion mining from social media, also known as sentiment analysis, provides up-to-date information; the reason is the proliferation of social media at various social levels. Sentiment analysis can be extracted from social posts using machine learning algorithms or lexicon-based approaches. The increasing of people who use social media led to more data being produced continuously. These huge amounts of data require an efficient framework to handle them. Hadoop provides a software framework for distributed storage and processing of big data using the MapReduce programming model and Hadoop Distributed File System (HDFS). This paper will discuss different approaches for sentiment analysis of big data, especially on Hadoop, and display their strengths and weaknesses through a comparative study.