A machine learning approach to predict academic performance based on student's regular activities.

Abstract

The Student Performance Prediction System uses machine learning to help predict student success based on factors like demographics, test preparation, and behavioral habits. Along with students reading and writing scores, the system seeks to provide educators and managers insightful analysis of data including gender, color, parental education level, and lunch type that can assist identify students who might require more support before academic issues get more intense. For the project, ridge regression was selected as it offers a nice mix between still producing accurate predictions and simplicity of understanding. Reliable predictions produced by the system were evaluated and may be applied to guide decisions and interventions in actual learning environments. Following significant data protection rules like GDPR and FERPA, we also ensured the system upholds students' privacy. Although the present version offers insightful analysis, we intend to enhance the system by adding additional data, investigating more sophisticated machine learning approaches, and thus improving its general accuracy. In the end, this technique is meant to provide a more customized and fair learning environment, so enabling kids to flourish and so preventing undetected falling behind.