Scholarlink

<p>Predicting student academic performance is essential for enhancing the quality of education and enabling early intervention strategies. This paper presents a framework for predicting student grades and marks using deep learning (DL) in intermediate and secondary education. The proposed framework incorporates data preprocessing, feature extraction, feature selection, and prediction to achieve greater accuracy and reliability. The dataset used in this study is based on students’ previous academic records and includes both qualitative and quantitative factors, such as demographic attributes, attendance records, past grades, and socio-economic variables. The collected data are preprocessed to address missing values, outliers, and inconsistencies. Improved Principal Component Analysis (IPCA) is employed for feature extraction, while an Improved Single Candidate Optimizer (ISCO) is used to select the most relevant features. The prediction model is developed using an Improved Convolutional Neural Network (ICNN) to ensure high accuracy and strong generalisation capability. Experimental results show that the proposed model achieves superior predictive accuracy with minimal error rates compared to existing methods. The findings confirm the effectiveness of DL-based approaches for forecasting student performance and provide practical insights that enable educators to identify at-risk students and implement timely interventions. Furthermore, this study contributes to the educational data mining literature by presenting an effective predictive framework that supports academic planning, resource allocation, and improved educational outcomes. Experimental evaluation shows that the proposed ICNN model achieves high predictive accuracy, with an R² score of 0.982 and low error values (RMSE = 0.45, MSE = 0.20), outperforming conventional machine learning and deep learning models.</p>

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

Enhanced academic performance prediction using ICNN and ISCO: a deep learning approach for intermediate and secondary education

  • 期刊论文

  • 发表日期：

    2026年3月10日

查看全文
  • S. Selvi,
  • M. Priya

摘要

Predicting student academic performance is essential for enhancing the quality of education and enabling early intervention strategies. This paper presents a framework for predicting student grades and marks using deep learning (DL) in intermediate and secondary education. The proposed framework incorporates data preprocessing, feature extraction, feature selection, and prediction to achieve greater accuracy and reliability. The dataset used in this study is based on students’ previous academic records and includes both qualitative and quantitative factors, such as demographic attributes, attendance records, past grades, and socio-economic variables. The collected data are preprocessed to address missing values, outliers, and inconsistencies. Improved Principal Component Analysis (IPCA) is employed for feature extraction, while an Improved Single Candidate Optimizer (ISCO) is used to select the most relevant features. The prediction model is developed using an Improved Convolutional Neural Network (ICNN) to ensure high accuracy and strong generalisation capability. Experimental results show that the proposed model achieves superior predictive accuracy with minimal error rates compared to existing methods. The findings confirm the effectiveness of DL-based approaches for forecasting student performance and provide practical insights that enable educators to identify at-risk students and implement timely interventions. Furthermore, this study contributes to the educational data mining literature by presenting an effective predictive framework that supports academic planning, resource allocation, and improved educational outcomes. Experimental evaluation shows that the proposed ICNN model achieves high predictive accuracy, with an R² score of 0.982 and low error values (RMSE = 0.45, MSE = 0.20), outperforming conventional machine learning and deep learning models.