Exploring the association patterns of student performance in depth can help administrators and teachers optimize the curriculum structure and teaching plans more specifically to improve teaching effectiveness in a col...Exploring the association patterns of student performance in depth can help administrators and teachers optimize the curriculum structure and teaching plans more specifically to improve teaching effectiveness in a college undergraduate major.However,these high-dimensional time-varying student performance data involve multiple associated subjects,such as student,course,and teacher,which exhibit complex interrelationships in academic semesters,knowledge categories,and student groups.This makes it challenging to conduct a comprehensive analysis of association patterns.To this end,we construct a visual analysis framework,called MAPVis,to support multi-method and multi-level interactive exploration of the association patterns in student performance.MAPVis consists of two stages:in the first stage,we extract students’learning patterns and further introduce mutual information to explore the distribution of learning patterns;in the second stage,various learning patterns and subject attributes are integrated based on a hierarchical apriori algorithm to achieve a multi-subject interactive exploration of the association patterns among students,courses,and teachers.Finally,we conduct a case study using real student performance data to verify the applicability and effectiveness of MAPVis.展开更多
目的针对飞机货舱配载方案评估中多属性决策的复杂性,以及现有评估方法在权重确定上过度依赖专家经验导致的主观偏差,或单纯依赖客观数据忽略决策者偏好的局限性,提出一种融合主观先验与客观数据驱动的混合赋权评估模型,以提供更为合理...目的针对飞机货舱配载方案评估中多属性决策的复杂性,以及现有评估方法在权重确定上过度依赖专家经验导致的主观偏差,或单纯依赖客观数据忽略决策者偏好的局限性,提出一种融合主观先验与客观数据驱动的混合赋权评估模型,以提供更为合理、可靠的配载方案择优决策支持。方法首先,引入大型语言模型(Large language model,LLM),构建“虚拟专家委员会”,通过精心设计的提示词工程,获取多维度、多情境下的主观权重。其次,针对传统熵权法对数据分布敏感、难以有效区分指标优劣等问题,提出一种改进的数据预处理熵权法(Improved data preprocessing entropy weighting method,IDPEW),该方法结合指标值的辨识度和信息熵的均衡性来确定客观权重。最后,将LLM生成的主观权重与IDPEW计算的客观权重进行加权组合,构建综合评价函数,对飞机货舱配载方案进行全面评估和排序。结果实验结果表明,LLM模拟专家意见时最关注“装载率”(主观权重0.2250),而IDPEW方法从数据中识别出“横向不平衡度”最具区分力(客观权重0.2481)。混合赋权模型(α=0.5)有效平衡了主客观偏好,在24个方案中精准识别出综合性能最优的方案,验证了模型在复杂情境下的稳定性。结论创新性地利用LLM低成本构建“虚拟专家”获取先验知识,并通过耦合指标辨识度与均衡性的IDPEW方法,提升了客观赋权精度。该模型克服了单一赋权的局限,为飞机货舱配载方案的科学评估提供了一种兼具可解释性和实用性的新范式。展开更多
基金supported by the National Natural Science Foundation of China(No.60873093).
文摘Exploring the association patterns of student performance in depth can help administrators and teachers optimize the curriculum structure and teaching plans more specifically to improve teaching effectiveness in a college undergraduate major.However,these high-dimensional time-varying student performance data involve multiple associated subjects,such as student,course,and teacher,which exhibit complex interrelationships in academic semesters,knowledge categories,and student groups.This makes it challenging to conduct a comprehensive analysis of association patterns.To this end,we construct a visual analysis framework,called MAPVis,to support multi-method and multi-level interactive exploration of the association patterns in student performance.MAPVis consists of two stages:in the first stage,we extract students’learning patterns and further introduce mutual information to explore the distribution of learning patterns;in the second stage,various learning patterns and subject attributes are integrated based on a hierarchical apriori algorithm to achieve a multi-subject interactive exploration of the association patterns among students,courses,and teachers.Finally,we conduct a case study using real student performance data to verify the applicability and effectiveness of MAPVis.
文摘目的针对飞机货舱配载方案评估中多属性决策的复杂性,以及现有评估方法在权重确定上过度依赖专家经验导致的主观偏差,或单纯依赖客观数据忽略决策者偏好的局限性,提出一种融合主观先验与客观数据驱动的混合赋权评估模型,以提供更为合理、可靠的配载方案择优决策支持。方法首先,引入大型语言模型(Large language model,LLM),构建“虚拟专家委员会”,通过精心设计的提示词工程,获取多维度、多情境下的主观权重。其次,针对传统熵权法对数据分布敏感、难以有效区分指标优劣等问题,提出一种改进的数据预处理熵权法(Improved data preprocessing entropy weighting method,IDPEW),该方法结合指标值的辨识度和信息熵的均衡性来确定客观权重。最后,将LLM生成的主观权重与IDPEW计算的客观权重进行加权组合,构建综合评价函数,对飞机货舱配载方案进行全面评估和排序。结果实验结果表明,LLM模拟专家意见时最关注“装载率”(主观权重0.2250),而IDPEW方法从数据中识别出“横向不平衡度”最具区分力(客观权重0.2481)。混合赋权模型(α=0.5)有效平衡了主客观偏好,在24个方案中精准识别出综合性能最优的方案,验证了模型在复杂情境下的稳定性。结论创新性地利用LLM低成本构建“虚拟专家”获取先验知识,并通过耦合指标辨识度与均衡性的IDPEW方法,提升了客观赋权精度。该模型克服了单一赋权的局限,为飞机货舱配载方案的科学评估提供了一种兼具可解释性和实用性的新范式。
