Akademik Məlumat İdarəetmə Sistemi
3rd World Conference on Communication and Computing, WCONF 2025, Raipur, India, 25 - 27 July 2025, (Full Text)
The use of auto-grading systems has become commonplace in higher education, particularly in computer science degrees, due to the rising demands of managing larger class sizes and providing students with quick and efficient feedback. By delivering scalable, consistent assessments, these technologies are vital in improving student performance evaluation. To gain a better understanding of how people are currently using it, they polled teachers from different schools to find out what they like, what they find difficult, and how they grade. According to the findings, the most popular systems provide significant levels of customization and integration, and output-based grading is also quite popular. By utilising advanced text preparation techniques as stemming, information extraction, tokenisation, and CNN, this study also suggests a hybrid auto-grading method that mixes ELM with CNN. In order to find trends and increase grading precision, information extraction compares student replies to instructor-provided answers, which is a data mining process. Outperforming current state-of-the-art approaches, the suggested DCNN-ELM model attained a precision level of 95.73%. These results show how intelligent auto-grading systems can help make assessments more reliable and efficient, which in turn can lead to better results for students and teachers in higher education.