Empirical Study and Comparison of Models via Multiclass Classification of COVID-19 Tweets using Natural Language Processing

Authors

  • Simran Anand Bachelor of Technology, Department of Computer Science Engineering, Vellore Institute of Technology, Amaravati, India
  • Deepasikha Mishra Assistant Professor, Department of Computer Science Engineering, Vellore Institute of Technology, Amaravati, India

Keywords:

Natural Language Processing, Sentiment Analysis, COVID-19, Machine Learning, Deep Learning, Data Analysis

Abstract

This paper represents the empirical study of sentiment analysis of the Covid-19 tweets during the pandemic period. Various industries have been psychologically affected throughout the country during this period. Through the Covid-19 tweets, a study has been established to determine whether the people’s attitude is positive, negative or neutral during the pandemic. In this work, Natural Language Processing, Exploratory Data Analysis and Machine Learning are used to analyze textual data consisting of Covid- 19 tweets. Different Machine Learning and Deep Learning techniques like Naive Bayes, Logistic Regression, Extreme Gradient Boost (XGBoost), Stochastic Gradient Descent, Random Forest, SVM, Bidirectional LSTMs (BiLSTM) and Backpropagation neural networks have been incorporated to analyze and predict efficiently. Finally, a comparison of each model’s performance based on evaluation metrics like accuracy, precision, recall and F1-score has been done.

Downloads

Download data is not yet available.

Downloads

Published

12-02-2022

Issue

Section

Articles

How to Cite

[1]
S. Anand and D. Mishra, “Empirical Study and Comparison of Models via Multiclass Classification of COVID-19 Tweets using Natural Language Processing”, IJMDES, vol. 1, no. 2, pp. 9–17, Feb. 2022, Accessed: May 20, 2024. [Online]. Available: https://journal.ijmdes.com/ijmdes/article/view/10