Prediction Model on Student Performance based on Internal Assessment using Deep Learning

Creative Commons License

Hussain S., Muhsin Z. F., Salal Y. K., Theodorou P., KURTOĞLU F., Hazarika G. C.

INTERNATIONAL JOURNAL OF EMERGING TECHNOLOGIES IN LEARNING, vol.14, no.8, pp.4-22, 2019 (ESCI) identifier identifier


Educational Data Mining and Deep learning play a crucial role in identifying academically weak students of an institute and help them by developing different recommendation systems to enhance their performance. These technologies direct the students for their future plan by discovering the precious hidden patterns from their history of information. Students from three colleges in Assam, India were considered in our research and their records were run on deep learning using the sequential neural model with the Adam optimization method. The study compared other classification methods such as the Artificial Immune Recognition System v2.0 and AdaBoost, to predict the results of the students. The highest classification accuracy achieved in this study was 95.34% produced by deep learning techniques. The Precision, Recall, F-Score, Accuracy, and Kappa Statistics Performance were calculated as a statistical decision to find the best classification methods. The dataset used in this study consisted of 10140 student records.