Smart Education Sentiment Analysis Combining CNN and ANN for Deeper Insights From Student Feedback
Usharani Bhimavarapu · Advances in computational intelligence and robotics book series · 2025
This study presents a comprehensive methodology for sentiment classification of student feedback data, utilizing a hybrid approach that combines Convolutional Neural Networks (CNN) and Artificial Neural Networks (ANN). The dataset consists of student feedback categorized into aspects such as teaching quality, course content, and campus facilities. The feedback is first preprocessed through tokenization, lowercasing, stop-word removal, and word embedding using pre-trained models like Word2Vec or GloVe. Aspect extraction is performed using BERTopic, with sentiment classification mapped to each aspect. The CNN is employed to extract local features, such as specific word patterns or n-grams that indicate sentiment, while the ANN captures the global context of the text to classify overall sentiment. Additionally, Part-of-Speech (POS) tagging is used to link aspect terms to corresponding sentiments, ensuring more precise sentiment mapping.