Research on English Teaching Assistant System Based on Convolution Neural Network with Spectral Attention Block

With the assistance of intelligent system, the English teaching and learning in information age required to carried out to change limitations of conventional English teaching assistant system. The Convolution Neural Network with Spectral Attention Block (CNN with SAB) is proposed for English teaching assistant system. The SAB produces significant frequency components in text and speech data which enhances the model's capability to extract linguistic features. It helps to filter out the noisy and irrelevant frequency components which leads to better performance. The SAB is integrated into CNN which enables efficient temporal and spectral feature fusion thereby enhance accuracy. the dataset is collected from student and it is preprocessed by data cleaning which address the missing values, duplicate data and inconsistency. The assessment metrics applied for estimating CNN with SAB performance are precision, recall, f1-score and accuracy. The CNN with SAB approach got 98.37% precision, 98.26% recall, 98.31 % f1-score and 99.13% accuracy which is better than state-of-art methods.