Jia Hongchun
College of Foreign Languages, Pingdingshan University, Pingdingshan, Henan, China.
PeerJ Comput Sci. 2024 May 31;10:e2074. doi: 10.7717/peerj-cs.2074. eCollection 2024.
In hybrid English teaching, there are many courses and various kinds of assessment, which put higher requirements for teachers' accurate and objective curriculum evaluation. This article adopts the clustering method of unsupervised learning to adapt to more data and give the evaluation method a specific generalization ability. A curriculum evaluation system based on AHP and clustering is proposed. Through hierarchical analysis values of online and offline average grades and final offline assessment scores, multiple hierarchical analysis is carried out, and the K-means method is adopted to refine course evaluation, and non-iterative calculation is carried out for non-deterministic numerical data to complete the final assessment of grades. Based on the sample test of the school's data in recent years, this article finds that the proposed method can distinguish different categories of students well, and the absolute error of K-means classification is less than 0.5. The proposed method can ensure the accurate evaluation of colleges and universities and reduce teachers' burden.
在混合式英语教学中，存在众多课程和各类评估，这对教师准确、客观的课程评价提出了更高要求。本文采用无监督学习的聚类方法以适应更多数据，并赋予评价方法特定的泛化能力。提出了一种基于层次分析法（AHP）和聚类的课程评价系统。通过对线上和线下平均成绩以及线下期末考试成绩进行层次分析赋值，开展多次层次分析，并采用K均值法细化课程评价，对不确定性数值数据进行非迭代计算以完成最终成绩评定。基于该校近年来数据的样本测试，本文发现所提方法能够很好地区分不同类别的学生，且K均值分类的绝对误差小于0.5。所提方法能够确保高校评价的准确性并减轻教师负担。
