<![CDATA[Reading comprehension is a critical skill for academic success; however, students with reading difficulties often face significant challenges in mastering this skill. This study aims to develop a deep learning-based error detection model for reading comprehension, integrated with Barrett’s Taxonomy as the analytical framework. The model is designed to automatically identify students’ error patterns through the analysis of reading comprehension test responses, thereby providing specific and timely feedback. The research method involves collecting reading test data from elementary school students, annotating errors based on Barrett’s Taxonomy categories, training the deep learning model for error classification, and testing the model’s validity and reliability. Preliminary results indicate that the model can accurately recognize different types of errors, which are then used as the basis for designing adaptive learning strategies tailored to each student’s error profile. These findings are expected to contribute to the development of literacy learning innovations that are responsive to individual needs while enriching artificial intelligence–based educational practices at the elementary school level.]]>
Copyrights © 2025
