Genetic Algorithm-Based Approach for Predicting Student Academic Success

A student's academic success and on-time graduation, with all the social and economic rewards that come with that, can be significantly improved by predicting a student's performance and the factors contributing to his academic success. There is a limit on how much time an academic adviser can devote to each student to detect curriculum problems, take the appropriate steps, and give the student guidance based on well-informed judgment. Determining whether students are at risk early in the program is essential to improving low-performing students' performance, retention, and completion rates. This will give academic advisers early signals of the need for intervention. In this research, we built and assessed a model based on genetic algorithms to forecast student performance and estimate a student's graduation grade point average (GP A). The model was assessed using a publicly available dataset created for machine learning techniques with R2 = 0.93, indicating it did well. The model was also applied to a dataset made from academic records of students who graduated with a bachelor's degree in computing from our institution; the input predictors were the students' grades in core information technology courses in addition to their high school average and the dependent variable was the graduation grade point average. According to our research, the best predictor of graduation success is student performance in the database management systems course, followed by software engineering, with networking and operating system courses having minimal bearing. Additionally, the findings indicate that only 54% of the graduation grade point average can be explained by the predictors used; hence, other academic and sociodemographic factors will need to be considered in future studies.