摘要
针对工程教育中难以实现个性化学习、难以提前识别和预警学生在未来课程中可能面临的困难等问题,提出基于模糊认知图(FCMs)的学生学习效果预测方法。首先,收集某高校通信工程专业215名学生期评成绩数据,并进行缺失值处理。然后,采用探索性因子分析方法构建“课程-毕业要求”支撑矩阵,基于Pearson相关性分析建立“课程-课程”相关性系数矩阵,并构建FCM初始邻接矩阵。接着,通过专家研讨会和非线性Heb‐bian学习算法改进模型,得到包含37个节点和271条因果权重关系的最终FCM。最后,通过4个场景验证此FCM的有效性。结果表明,该模型以学生前序课程目标达成情况为输入,能够预测后序课程目标和毕业要求达成的变化情况,提前识别学生潜在学习障碍,为教师和学生提供个性化教学依据,有效提升学生学习成效。
To address challenges in personalized learning and early identification of potential difficulties in future courses within engineering education,a fuzzy cognitive maps(FCMs)‑based method for predicting students’learning effect was proposed.First,end‑of‑term evaluation data from 215 Communications Engineering students were collected,processed for missing values,and analyzed using exploratory factor analysis to construct a“course‑graduate attributes”support matrix.Then,Pearson correlation analysis was applied to establish a“course‑course”correlation matrix,forming the initial FCM adjacency matrix.The final FCM,which comprises 37 nodes and 271 edges,was obtained through expert workshops and a nonlinear Hebbian learning algorithm.Finally,the validity of this FCM was verified through four scenarios.The results indicate that the model can predict changes in subsequent course objectives and graduate attribute attainment based on prerequisite course performance,enabling early identification of learning obstacles and providing personalized instructional support,ultimately enhancing students’learning achievement.
作者
旷怡
邓家俊
段斌
KUANG Yi;DENG Jiajun;DUAN Bin(School of Automation and Electronic Information,Xiangtan University,Xiangtan 411105,China;Xiangtan University Professional Certification Guidance Center,Xiangtan 411105,China)
出处
《东南大学学报(自然科学版)》
北大核心
2025年第4期1210-1216,共7页
Journal of Southeast University:Natural Science Edition
基金
国家自然科学基金资助项目(61379063)
湖南省普通高等学校教学改革研究重点资助项目(HNJG‑2022‑0079).
