Scholarlink

<p>As technology advances, traditional teaching, training, and performance evaluation approaches are being replaced by computer-based tools (CBT) such as online video courses, online exams, offline video classes, and flip classes. To improve the teaching and learning performance of the students, the education system (ES) assesses the course learning outcomes (CLO) by using the bloom taxonomy (BT). Owing to the abrupt COVID-19 epidemic, traditional classrooms are molded into online classes; therefore, it is very hard to assess the CLO because physical and online teaching pedagogies are heterogeneous, and it is very hard to classify either content has been generated by the students or any other online sources used to answer the questions. Thus, it is necessary to construct the classification model using artificial intelligence (AI) techniques and to assist educational stakeholders like schools, colleges, universities, NGOs, and the government in measuring student performance. This research proposes a comparative analysis of machine learning techniques to mitigate all the above-stated problems. It proposes methods to perform data cleaning in its first phase, whereas the second phase constructs classification models using machine learning algorithms such as decision tree, support vector machine, and random forest. In the third phase, performance evaluation is performed to estimate the results. The highest accuracy was achieved using a decision tree with 96.99% and 10&#xa0;k-fold cross-validation. This research contributes (a) a real-world dataset and (b) the highest classification accuracy to analyze education.</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

Comparison of machine learning techniques to predict students’ CGPA by using course learning outcomes datasets

  • 期刊论文

  • 发表日期：

    2025年4月28日

查看全文
  • Saima Siraj,
  • Fida Hussain Dahri,
  • Jamil Ahmed Chandio,
  • Akhtar Hussain Jalbani,
  • Asif Ali Laghari

摘要

As technology advances, traditional teaching, training, and performance evaluation approaches are being replaced by computer-based tools (CBT) such as online video courses, online exams, offline video classes, and flip classes. To improve the teaching and learning performance of the students, the education system (ES) assesses the course learning outcomes (CLO) by using the bloom taxonomy (BT). Owing to the abrupt COVID-19 epidemic, traditional classrooms are molded into online classes; therefore, it is very hard to assess the CLO because physical and online teaching pedagogies are heterogeneous, and it is very hard to classify either content has been generated by the students or any other online sources used to answer the questions. Thus, it is necessary to construct the classification model using artificial intelligence (AI) techniques and to assist educational stakeholders like schools, colleges, universities, NGOs, and the government in measuring student performance. This research proposes a comparative analysis of machine learning techniques to mitigate all the above-stated problems. It proposes methods to perform data cleaning in its first phase, whereas the second phase constructs classification models using machine learning algorithms such as decision tree, support vector machine, and random forest. In the third phase, performance evaluation is performed to estimate the results. The highest accuracy was achieved using a decision tree with 96.99% and 10 k-fold cross-validation. This research contributes (a) a real-world dataset and (b) the highest classification accuracy to analyze education.