Deep Neural Network Model for Predicting Foreign Language Classroom Anxiety through Eye-Tracking Data

The present study introduced an innovative approach utilizing a Deep Neural Network (DNN) to automatically predict Foreign Language Classroom Anxiety (FLCA) levels among foreign language learners. The proposed model incorporated data obtained from the Short-form Foreign Language Classroom Anxiety Scale (S-FLCAS) and eye-tracking measures to develop an efficient real-time prediction system of FLCA levels. A sample of 110 master's degree students from the Higher Institute of Management of Gabes, Tunisia, participated, contributing to the collection of data crucial for training and testing the DNN-based model. The DNN model was trained and tested using this combined dataset, aiming to establish patterns and correlations between the reported anxiety levels and the corresponding eye movement data. The findings demonstrated promising results, indicating the potential of the DNN-based model in objectively assessing FLCA levels, thereby helping teachers minimize emotional barriers to language acquisition and enabling timely intervention when elevated anxiety levels are identified. Overall, this work presents a pioneering methodology that combines psychometric assessments and machine learning techniques, showing how computational approaches can be leveraged to model and predict FLCA, thereby contributing to the broader field of computer engineering applications in education and intelligent learning systems.