<![CDATA[This study aims to analyze language anxiety among students in Arabic learning under the Merdeka Curriculum and to apply a deep learning approach to detect and predict the factors causing such anxiety. The research method employed is descriptive qualitative, with data collected through in-depth interviews, classroom observations, and documentation analysis at MTs Nurul Islam Mojokerto. The findings indicate that language anxiety is closely related to the fear of making mistakes in pronunciation and sentence structure, as well as the lack of speaking practice in more informal contexts. However, students who demonstrate greater self-confidence in speaking tend to experience lower levels of anxiety. The application of deep learning in this research shows that technology can identify anxiety patterns based on students’ interactions with speaking and writing tasks. In light of these findings, it is recommended that Arabic learning place stronger emphasis on flexible speaking practices and the use of deep learning tools to analyze students’ anxiety. The study also suggests further research to develop more specific deep learning models to address language anxiety across different levels of education.]]>
Copyrights © 2025
