Enhancing student performance prediction through stream analysis dataset using modified XGBoost algorithm

Abstract

In the education domain, predicting the academic performance of students has become an essential task to improve learning outcomes. In this study, we propose a Modified XGBoost (MXGB) model for predicting student performance using stream-based analysis of the dataset. We used a modified version of the XGBoost algorithm using cross-validation, which incorporates stream-based analysis to enhance its performance on real-time data. We preprocessed the dataset and applied feature engineering techniques to extract relevant features for building the model. We trained the MXGB model on the preprocessed dataset and evaluated its performance using various metrics such as accuracy, precision, sensitivity, and F1-score. The results show that our model outperforms the baseline XGBoost model and achieves high accuracy in predicting the student's academic performance. Our model can assist educational institutions in identifying students who are at risk of performing poorly and providing them with timely intervention to improve their academic outcomes.