An Enhanced Sentiment Analysis Based Teaching And Learning Process Via Audio Feedback Using MFCC & PSO

Accurately gauging student mood in the rapidly evolving digital education landscape is particularly challenging, given the limitations of conventional evaluation techniques in virtual learning environments. Meaningful improvements in curriculum, instructional materials, pedagogical approaches, and assessments necessitate the proper evaluation of student feedback. This research integrates a Naïve Bayes classifier with Particle Swarm Optimization (PSO) to enhance the accuracy of sentiment analysis in student feedback. The proposed method addresses the challenges of feature extraction, feature selection, and noisy data, such as background noise, variations in speech, and inconsistencies in audio quality—all crucial for effective sentiment analysis categorization. Student feedback is collected in audio format, with features extracted using Mel-Frequency Cepstral Coefficients (MFCC). By optimizing feature selection, the PSO method improves the Naïve Bayes classifier's accuracy. The sentiment analysis report categorizes feedback as positive, negative, or neutral, giving faculty members concrete information to improve their teaching tactics. This strategy fosters a customized and encouraging learning environment that is beneficial to each student's progress by guaranteeing a deeper comprehension of the emotions and thoughts of students.