An Improved Vectorization-Based Emotion Detection Using Tuned Inverse Document Frequency Approach
R. Vanitha · International Journal of Electronics and Communication Engineering · 2024
Emotion analysis of social media content has garnered significant attention due to its potential to reveal valuable insights into people’s feelings and opinions. This study is motivated by the need to understand better the emotions individuals express when posting their views on social media. The objective is to explore and compare the effectiveness of two machine learning methods, a Twin Support Vector Machine (TWSVM) and a novel approach called Tuned Inverse Document Frequency (TUNED-IDF) vectorizer, in accurately detecting emotions from textual data. To achieve this objective, the research process involves first applying the TWSVM algorithm, which examines the factors influencing emotions and their connection to the dependent variable. Next, our innovative TUNED-IDF vectorizer converts the textual content into numerical representations, leveraging its properties to improve accuracy in emotion analysis. The findings of this study showcase the remarkable performance of the TUNED-IDF approach, achieving an impressive accuracy level of 94.4%, surpassing existing methods in emotion detection. By employing this method, the research successfully predicts people’s emotions with higher precision and efficacy than traditional machine learning models. The significance of this research lies in its contribution to the field of emotion analysis, particularly in the context of social media. Understanding the emotions conveyed in online communication is crucial for various applications, such as sentiment analysis, market research, and public opinion monitoring. The insights gained from this study offer valuable opportunities for better comprehension of individuals’ sentiments in the digital age and lay the groundwork for enhanced emotion analysis techniques.