Elevating educational insights: sentiment analysis of faculty feedback using advanced machine learning models

The article presents a comprehensive analysis of sentiment in faculty feedback provided by students using various machine learning approaches. Faculty feedback is crucial in shaping student learning experiences and academic outcomes. Through sentiment analysis, the paper aims to discern the emotional tone of feedback, gauging its positivity, effectiveness, or areas for improvement. This study explores methodologies to classify feedback sentiments employing five machine learning models: “Support Vector Machine (SVM), Random Forest (RF), Stochastic Gradient Descent (SGD), Multilayer Perceptron (MLP), and Multinomial Naive Bayes (MNB).” A total of 5000 engineering graduate responses were processed through TF-IDF feature extraction, which converts textual information into numeric forms for analysis. The evaluation of the models was conducted by measuring their performance through accuracy, precision, recall and F1-score metrics. Further, comparing these metrics, the study identifies the best-performing machine learning approach for sentiment analysis on faculty feedback. The analysis aids in understanding student perceptions and offers actionable insights for educators to enhance their feedback practices, fostering a more positive and conducive learning environment. The results demonstrate that the Random Forest model performs the best, achieving an accuracy of 91%, precision of 94%, recall of 85%, and an F1 score of 89%. These findings suggest that sophisticated models like Random Forest are preferable for accurate and reliable sentiment analysis in educational settings. Such insights can guide educational institutions in leveraging machine learning technologies to improve feedback evaluation processes and ultimately enhance the overall educational experience.