Twitter Sentiment Analysis using Machine Learning
Abdullah Ikram, Mahesh Kumar, Geetika Munjal
Abstract
In this exponentially evolving age of technology, the emergence of social media has enveloped all life forms on our planet. Twitter is that social media platform on which millions of users post their opinions or sentiments as tweets daily on a plethora of topics. Sentiment analysis is a technique for analyzing source material, determining its sentiment, and categorizing it as favorable, negative, or unbiased. Twitter Sentiment Analysis is the utilization of several libraries in order to collect data from the Twitter API on a topic, assess it, and derive public and private views on it. It is of immense importance for a corporation to employ opinion mining to examine its online reputation for maximizing its profits. This study aims to acquire and transform the dataset named Sentiment140 from an unstructured to a structured format through data preprocessing, perform feature selection on the dataset, train, evaluate, and compare various machine learning models using it, and utilize the classifier that achieves the highest accuracy along with vectorization through a pipeline to identify the sentiment of a new tweet given as an input.