Simulating and Predicting Students’ Academic Performance Using a New Approach based on STEAM Education

In many countries, particularly in Iraq, the students’ academic performance (SsAP) system is based on final grade scores high school. This school may not reflect intelligence level or interests that link student to a relevant university. Also, skills are used predict score-related college. this research, seven-subject, one-grade, one-output (SOO) model was proposed simulate classic SsAP show predicting completely previous year’s score and skills. Moreover, twelve-year, seven-output (STS) model, which seven parallel deep neural networks with scaled conjugate learning algorithm, employed determine science, technology, engineering, art, mathematics (STEAM) across 12 grades their corresponding most appropriate article contributed constructing two models: SOO simulates classical Iraqi education system, STS predicts acceptance of students according STEAM what makes it different from research. The results revealed properly simulated system. Furthermore, new approach successfully predicted line over twelve-year period. overall accuracy rate models (SOO STS) about 99% 10-5 histogram errors between target actual output. However, optimized epochs 1000 while got 10–600 epochs.