Effectiveness of Logistic Regression for Sentiment Analysis of Tweets About the Metaverse

In the evolving landscape of digital communication, sentiment analysis provides crucial insights into public opinions, particularly concerning emerging technologies like the Metaverse. While various models have been employed to perform sentiment analysis, there is a need to assess the effectiveness of traditional machine learning approaches, specifically Logistic Regression (LR), given its advantages in terms of simplicity and interpretability. This study employs LR to analyze a dataset of tweets related to the Metaverse, focusing on preprocessing techniques such as tokenization and vectorization to optimize model performance. The LR model demonstrated high efficacy, achieving an accuracy of 96% with precision, recall, and F1-scores of 0.93, 0.96, and 0.95 respectively for negative sentiments, and 0.97, 0.95, and 0.96 for positive sentiments. The findings suggest that Logistic Regression remains a robust tool for sentiment analysis in social media contexts, offering significant implications for businesses and developers interested in the public perception of new technologies such as the Metaverse.