NLP and Machine Learning for Sentiment Analysis in COVID-19 Tweets: A Comparative Study

In response to the COVID-19 pandemic, a novel technique is given for assessing the sentiment of individuals using Twitter data obtained from the UCI repository. Our approach involves the identification of tweets with a discernible sentiment, followed by the application of specific data preprocessing techniques to enhance data quality. We have developed a robust model capable of effectively discerning the sentiments behind these tweets. To evaluate the performance of our model, we employ four distinct machine learning algorithms: logistic regres sion, decision tree, k-nearest neighbor and BLSTM. We classify the tweets into three categories: positive, neutral, and negative sentiments. Our performance evaluation is based on several key metrics, including accuracy, precision, recall, and F1-score. Our experimental results indicate that our proposed model excels in accurately capturing the perceptions of individuals regarding the COVID-19 pandemic.