Student Performance Prediction based on Novel Clustered Regression

The ability to forecast the academic achievement of students is essential for institutions of higher learning. An educational institution can predict students' academic development by using data mining techniques. To predict the achievement of students in education, this paper proposes cluster-based regression approaches and investigates the value of a student achievement dataset. The efficiency of cluster-based regression (CR) algorithms is demonstrated by the results of our experiments on numerical data sets. The proposed CR technique aims to find the cluster for a new data object with the lowest processing cost. Future research will focus on developing an optimal model (F1-Score:0.89) that utilizes domain expertise and explores various methods for solving multi-instance prediction problems.