More openness and accountability from the higher education industry is being demanded, in part, because people are worried about how well universities are using their budgets to improve student learning. There is a dearth of information regarding how financial aid affects students’ performance in the classroom. The purpose of this research was to look at how several variables, such as student engagement and certain self-reported learning outcomes, relate to educational spending. College Board and Integrated Postsecondary Education Data System (IPEDS) institutional data were combined with student responses from the 2014 National Survey of Student Engagement (NSSE) to accomplish this. Institutional and student variables were also accounted for in the system. During the preprocessing, feature selection, and model training phases, there is an absolute no-break condition. One part of data preparation is checking for missing values; another part is encoding and normalizing the data. After the relative features were developed, the proposed approach used the random forest method’s feature selection to find out what qualities separate students who didn’t finish the course from those who did. Training a GA-BP-ANN model requires feature extraction. With the data that is already accessible, the suggested approach outperforms the current best practices. An improvement in accuracy of 96.41% was achieved by the results.