A Hybrid Machine Learning Model for Grade Prediction in Online Engineering Education

Facing the disruption caused by COVID-19 pandemic, emergence of imposed and exclusive online learning revealed challenges for researchers worldwide, as reforming curricula shortly collecting data accumulated monitoring stu-dents’ commitment academic performance. With this pool data, research explores grade prediction in a first-semester mechanical engineering CAD module, after testing performance reform specific curricula. A hybrid model has been created, based on 35 variables having filtered out statistical analysis shown to be strongly correlated students’ specif-ic module during first semester year 2020-2021. The hy-brid consists Generalized Linear Model. It’s fitting errors are used an extra predictor artificial neural network. architecture network can described following sizes: size input layer (36), hidden (1) output (1). Since new factors affect achievements, trained 70% participants predict remaining 30%. therefore divided into three subsets, with training set sample one predicting test (15%) validation (15%). final form resulted coefficient determination equal 1 (R = 1). This means that process 100% success rate, terms associating independent dependent variable (grade).