Educational Data Mining for Predicting Academic Attrition in Iraqi Universities: A Hybrid KDD-Based Approach using Weighted Euclidean Distance and Neural Architectures

Academic attrition within the higher education sector of Iraq is accepted as a notable phenomenon with substantial socio-economic impacts. Within the digital revolution environment of the higher education sector of Iraq, there is a growing need to exploit the student data repositories as a strategic tool. The objective and purpose of this research is to enhance the existing strong framework, based on the Knowledge Discovery in Databases (KDD) process, whereby the risk factors associated with academic attrition can be proactively identified. Through the use of advanced image mining techniques, such as a texture-like matrix, along with a precise prediction model utilizing a Weighted Euclidean Distance with parameters relevant to education, a precise prediction model is now attainable. A comprehensive evaluation of different architectures for machine learning was carried out using a rich dataset of 5,000 students, designed based on the Iraqi context for the educational system. The results show the benefits of using Artificial Neural Network (ANN) methodology with Knowledge Discovery in Databases (KDD) methods, resulting in an accuracy rate of 94.21%. The research provides a strategic framework for Iraqi university administrators seeking to improve student retention using early warning systems