Machine learning-based sentiment analysis of Twitter data

Abstract

The paper analyzes the views of Twitter users on the COVID-19 corona virus pandemic based on machine learning algorithms. The role of sentiment analysis increased with the advent of the social network era and the rapid spread of microblogging applications and forums. Social networks are the main sources for gathering information about users’ thoughts on various themes. People spend more time on social media to share their thoughts with others. One of the themes discussed on social networking platforms Twitter is the COVID-19 corona virus pandemic. In the paper, machine learning methods as Naive Bayes (NB), Support Vector Machine (SVM), Random Forest (RF), Neural Network (NN) are used to analyze the emotional “color” (positive, negative, and neutral) of tweets related to the COVID-19 corona virus pandemic. The experiments are conducted in Python programming using the scikit-learn library. A tweet database related to the COVID-19 corona virus pandemic from the Kaggle website is used for experiments. The RF classifier shows the highest performance in the experiments.