Sentiment Analysis of Tweets on Social Issues Using Machine Learning Approach

Abstract

Since the arrival of Web 2.0, there has been a growing interest in knowing the opinions of Internet users who express themselves spontaneously and in real time. This mass of opinion data is accessible with web mining tools, with a constantly renewed collection of information. Sites have specialized in collecting these opinions in certain fields (movie reviews, for example), and Internet users have become accustomed to consulting the opinions and ratings submitted by others as soon as they have to make a purchasing decision for a technical product, or even for a hotel reservation. Opinions are therefore of interest to Internet users and have given rise to multiple applications and services, which creates a virtuous circle of encouragement to give one's opinion and even to be recognized as giving relevant opinions and followed by others. But this data is also of interest to brands and research firms who are trying to understand this "aggregated crowd sentiment". Often sensitive to the "your reputation can be destroyed because of a blog comment" fantasy, brands are concerned about their online identity but also seek to better understand the expectations and criticisms that Internet users have of them. Hence the growing development of techniques to capture these evaluations from Internet users, ranging from the simple counting of positive or negative comments to a more detailed analysis of the content of these comments. The purpose of this document is to provide detailed steps in the process of analyzing sentiments on a data using machine learning.