Machine Learning Algorithm For Sentimental Analysis of Twitter Feeds

Abstract A sentiment can be defined as a personal positive or negative feeling. Opinion mining is the computational technique for extracting data, classifying it then understanding, and assessing the opinions expressed in various contents. Huge amount of data is generated daily on various social networking sites. Millions of people are posting their likes, dislikes, comments about anything daily on social networking sites. This paper discusses an approach where a publicized stream of tweets from the Twitter microblogging site are preprocessed and classified based on their emotional content as positive, negative and neutral and algorithm which is used to classify these sentiments. Algorithm performance is improved by reducing words in tweet to their root form through mechanism of pre-processing before passing them to sentiment analyzer. Hence, the algorithm classifies tweets as neutral, positive or negative with respect to a query term. This is very useful for the companies and other organizations who want to know the people’s opinion about their products or the customers who want to get the feedback from others about product before purchase or also for election exit polls.