Sentiment Analysis of Covid-19 Tweets using Various Machine Learning Techniques with Hyperparameter Tuning on Twitter Database

Abstract

People around the world are now being affected by the 2019 coronavirus disease (COVID-19) epidemic. People are using social media to express their opinions and general thoughts about the epidemic, which has affected their daily lives both in general and during lockdown periods. Twitter, one of the most used social media platforms, has seen a massive surge in tweets related to the coronavirus, including happy, sad, fear, and anger tweets, in a very short period. In the COVID-19 tweet dataset, there is a total number of tweets count is 179108 data in unstructured form, and after that preprocess that data to become semi structured form. After preprocessing feature extraction is used by applying CountVecotorizer and TF-IDF methods. Various Machine Learning Models such as Random Forest (RF), Decision Tree (DT), Support Vector Machine (SVM), and Stochastic Gradient Decent (SGD) are considered machine learning methods for sentiment analysis, To improve performance (accuracy), with Hyperparameters tuning method is used.