On Students’ Sentiment Prediction Based on Deep Learning: Applied Information Literacy

The main purpose of this study is to understand the assessment of students’ sentiments on improving the role of library service quality after the pandemic and deduce the trends consistent with the new phase of technology. In this study, convolution bidirectional learners (CBLs) are employed for evaluating such as Convolution Bidirectional Long Short-Term Memory (CBLSTM), Gated Recurrent Unit (CBGRU), Simple Recurrent Neural Network (CBSRNN), Combination of Convolutional Bidirectional LSTM+GRU (CBLSTMGRU), SimpleRNN+LSTM (CBSLSTM), SimpleRNN+GRU (CBSGRU), Attention, and Transformer TFBERT model. The results show that CBSLSTM (94.82% accuracy) and CBLSTMGRU (94.78% accuracy) surpassed the other models in predicting Vietnamese students’ sentiments.