Determining Twitter Trending Hashtags and Sentiments Associated via Machine Learning Approaches

Social media platforms such as Twitter serve as a powerful tool for real-time information dissemination and worldwide communication, shaping public opinions and providing a platform for diverse voices to be heard. The current study proposes methods to identify trending hashtags on Twitter over a six-month period (June to October 2021) and analyze the sentiment dynamics linked to these hashtags. In order to carry out trend analysis three classification algorithms—Naive Bayes, Gradient Boosting Decision Tree, and Random Forest—were evaluated. The results indicate that the Random Forest model outperformed the other two algorithms in accurately identifying trending hashtags. In addition to this, a comprehensive sentiment analysis was conducted using both rule-based and lexicon-based approaches. The VADER sentiment analysis was utilized as the foundation for the Support Vector Machine classifier which achieved an accuracy of 91.84%, while the TextBlob library provided additional sentiment insights. The majority of research results, leaned towards neutrality, indicating challenges in capturing sentiment complexities present in Twitter conversations. The findings highlight the effectiveness of machine learning approaches, while also recognizing the potential for deep learning techniques to enhance sentiment analysis in future research.