<![CDATA[This study discusses the prediction of student performance by considering factors that can influence academic performance. In this research, the SelectK-Best feature selection technique and linear regression were used to enhance the accuracy of the prediction. The selection of this topic is based on the importance of understanding the factors that influence student performance and how feature selection can help build more efficient models. The methods applied in this study include data exploration through EDA, the use of SelectK-Best to select the most significant features, and linear regression to build the prediction model. The evaluation metrics show that the model with feature selection achieved MAE of 0.6293, MSE of 0.5945, RMSE of 0.7711, and R² Score of 0.9144, demonstrating the model's excellent performance. In contrast, the model without feature selection did not produce better results than the model with feature selection. This emphasizes the importance of applying feature selection techniques in building more accurate prediction models. This study contributes to predicting student performance through the use of systematic and effective methods, while also opening opportunities for further research in the context of education and more diverse data.]]>
Copyrights © 2025
