Design and Implementation of an Intelligent Tutoring System Based on Deep Learning in Computer Application Courses

In response to the persistent challenges in computer application courses at application-oriented universities-such as significant disparities in students' foundational knowledge, the lack of personalized tutoring, and the substantial workload imposed on instructors-this study presents the design and implementation of an intelligent tutoring system (ITS) driven by deep learning technologies. The system integrates advanced Transformer and recurrent neural network (RNN) models as its core analytical engines to comprehensively assess students' learning patterns, programming assignments, Q&A interactions, and behavioral learning logs. By leveraging this rich dataset, the ITS can accurately identify individual knowledge gaps, pinpoint specific learning difficulties, and deliver highly tailored educational support. The system offers customized learning strategies, targeted code correction prompts, adaptive practice recommendations, and real-time guidance that aligns with each student's learning pace and proficiency level. Experimental evaluations indicate that students engaging with the ITS exhibit notable improvements in programming competence, problem-solving efficiency, and independent learning capabilities, while instructors benefit from a significant reduction in direct tutoring time and workload. Furthermore, the system enhances the overall teaching quality by promoting precision-oriented instruction and efficient tutoring practices, establishing a scalable and practical model for intelligent, data-driven education in computer courses at applied undergraduate programs. This approach not only addresses the heterogeneity of student knowledge but also facilitates continuous monitoring and adaptive guidance, thereby fostering a more effective and personalized learning environment.