Abstract

Nowadays, the prediction of student performance is still complicated to analyze the talent of individuals and the effort to improve their academic performance. Moreover, the researchers are performed to analyze the outcomes of student performance but the educational database consists of a huge data volume, which is hard to train the small sample. In this research work, a new hybrid deep learning model with optimized entropy rough set theory is developed to predict the student’s performance accurately. The preprocessing phase is performed with outlier removal and the data-filling method. The features are mined from the preprocessed data by Entropy weighted Rough set-based feature mining. A novel meta-heuristic hybrid Galactic Rider Swarm Optimization (GRSO) algorithm is developed for feature mining. A hybrid deep learning-based Convolutional Recurrent Network (CRN) is implemented for prediction, where the classification performance is improved by the GRSO algorithm. Here, the hyperparameters of the Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) are optimized by the GRSO algorithm. Here, the sensitivity and accuracy rate of the recommended GRSO-CRN method attain 94% and 93%. The simulation outcome of the proposed GRSO-CRN model achieves enriched performance.

摘要

如今，学生成绩的预测仍然很复杂，无法分析个人的才能以及为提高学业成绩所做的努力。此外，研究人员被执行以分析学生表现的结果，但教育数据库包含巨大的数据量，难以训练小样本。在这项研究工作中，开发了一种新的具有优化熵粗糙集理论的混合深度学习模型来准确预测学生的表现。预处理阶段通过异常值去除和数据填充方法执行。这些特征是通过基于熵加权粗糙集的特征挖掘从预处理数据中挖掘出来的。开发了一种新的元启发式混合银河骑士群优化 (GRSO) 算法用于特征挖掘。实现了基于混合深度学习的卷积递归网络 (CRN) 进行预测，其中 GRSO 算法提高了分类性能。这里，卷积神经网络（CNN）和递归神经网络（RNN）的超参数通过GRSO算法进行了优化。这里，推荐的 GRSO-CRN 方法的灵敏度和准确率达到 94% 和 93%。所提出的 GRSO-CRN 模型的仿真结果实现了丰富的性能。