Predicting Emotions from Twitter Posts: A Comparative Study of Machine Learning Methods
Peihang Li · Advances in computer science research · 2023
With the increasing importance of social media platforms such as Twitter, understanding the emotions expressed in text data has become crucial for various applications.Manual analysis of the vast amount of user-generated content is impractical, highlighting the need for automated classification techniques.This study focuses on evaluating different machine learning methods for predicting emotions from Twitter posts, specifically examining Multinomial Naive Bayes (MultinomiaNB), Support Vector Machines (SVM), and the Random Forest.A dataset containing over 4000 labeled tweets, categorized as positive, neutral, or negative, is used for evaluation purposes.The challenges associated with predicting emotions from Twitter text, including natural language ambiguity and noise, are carefully considered.The results demonstrate that all models perform well, with SVM exhibiting a slight advantage.This study contributes to a deeper understanding of user emotions and public opinion in social media contexts.Future research directions include refining preprocessing techniques, exploring advanced methods like deep learning, incorporating additional features, and leveraging ensemble learning approaches in order for higher accuracy.