Neural Network-Based Sentiment Analysis for Real-Time E-Learning Feedback and Content Adjustment

The increasing trend of e-learning, requires intelligent systems which dynamically adapt content based on learner feedback. Traditional approaches to sentiment analysis, are unable to cope with the contextual variations in usergenerated texts, rendering them ineffective for real-time applications. The proposed method RNN is used for E-learning for contextual understanding is augmented by utilizing the e-learning-specific benchmarking dataset along with data from social media. After that data pre-processing is done by noise removal, removing common words and data splitting. Then RNN architecture is used that takes sequential data as input. The python is used to implement the proposed method. Comparing, the experimental results indicates RNN achieves an accuracy of 98.3% that outperforms Bi-LSTM (98.5%), GRU (98.4%) and transformer (97.1%). The RNN proves to work better for real-time applications. This research demonstrates the applicability of deep learning in adaptive learning environments while proposing future enhancement incorporation of hybrid models with attention mechanisms.