Artificial Neural Networks and the Modelling and Prediction of Australian Students’ Academic Achievement

This project uses Artificial Neural Networks (ANN), a type of Machine Learning, to study the nonlinear relationships between educational resources and academic achievement proposed by the Actiotope Model of Giftedness (AMG). ANN outperforms Structural Equation Modelling, a benchmark linear model, for five out six measures of academic achievement and has similar performance for the sixth. ANN can predict improvements to academic achievement calculated from hypothetical increases in resources. These results confirm the presence of nonlinear relationships as hypothesised by the AMG and act as a proof-of-concept for the use of ANN to study other nonlinear relationships in education research.