Scholarlink

This study presents a comprehensive analysis of various regression models to predict student performance, focusing on the impact of preprocessing, feature selection, and model evaluation. The research leverages a dataset comprising key academic, demographic, and social factors affecting student outcomes. Initial preprocessing involved converting categorical variables into numerical formats using Label Encoding and One-Hot Encoding, ensuring the dataset was suitable for machine learning algorithms. Feature selection techniques and correlation analysis, were employed to identify the most significant predictors of student performance, optimizing model efficiency. Six regression models were evaluated: Linear Regression, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Decision Tree, Random Forest, and Gradient Boosting. Each model was trained on the selected features, and performance was assessed using standard regression metrics-Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R2 Score. The results revealed that Linear Regression achieved the highest R2 score and the lowest error metrics, outperforming the other models. While SVM and Gradient Boosting also demonstrated strong predictive capabilities, Decision Tree models showed signs of overfitting, and Random Forest, though effective, did not surpass the top models.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Comparative Analysis of Predictive Models for Analysing Demographics and Academic Features to Predict Student Performance Using Machine Learning Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Harshvardhan Tiwari,
  • Neel Pandey

摘要

This study presents a comprehensive analysis of various regression models to predict student performance, focusing on the impact of preprocessing, feature selection, and model evaluation. The research leverages a dataset comprising key academic, demographic, and social factors affecting student outcomes. Initial preprocessing involved converting categorical variables into numerical formats using Label Encoding and One-Hot Encoding, ensuring the dataset was suitable for machine learning algorithms. Feature selection techniques and correlation analysis, were employed to identify the most significant predictors of student performance, optimizing model efficiency. Six regression models were evaluated: Linear Regression, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Decision Tree, Random Forest, and Gradient Boosting. Each model was trained on the selected features, and performance was assessed using standard regression metrics-Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R2 Score. The results revealed that Linear Regression achieved the highest R2 score and the lowest error metrics, outperforming the other models. While SVM and Gradient Boosting also demonstrated strong predictive capabilities, Decision Tree models showed signs of overfitting, and Random Forest, though effective, did not surpass the top models.