面向复杂工程问题的多元统计分析教学改革——“问题链–证据链”模式的设计与实践

Teaching Reform of Multivariate Statistical Analysis for Complex Engineering Problems—Design and Practice of a “Problem Chain-Evidence Chain” Model

摘要: 面向新工科复杂工程问题培养需求，针对多元统计分析课程中重计算、轻判断，统计方法与工程问题衔接不足等问题，本文尝试构建“问题链–证据链”教学模式。课程以工程数据分析任务为主线，围绕主成分分析、因子分析、聚类分析、判别分析和偏最小二乘回归等内容组织教学，引导学生从数据现象出发选择方法、解释结果并判断结论边界。课程评价重点考查问题识别、方法选择、结果解释和证据判断，使学生逐步形成面向实际工程问题的数据分析与判断能力。

Abstract: In response to the requirements of Emerging Engineering Education for cultivating students’ ability to address complex engineering problems, this study develops a “problem chain-evidence chain” teaching model to address common issues in Multivariate Statistical Analysis, including an overemphasis on computation, insufficient attention to judgment, and weak connections between statistical methods and engineering problems. The course is organized around engineering data-analysis tasks and covers principal component analysis, factor analysis, cluster analysis, discriminant analysis, and partial least squares regression. Students are guided to identify problems from observed data patterns, select appropriate statistical methods, interpret model results, and determine the boundaries of the conclusions supported by the available evidence. Course assessment emphasizes problem identification, method selection, result interpretation, and evidence-based judgment, with the aim of gradually developing students’ ability to analyze engineering data and make informed judgments in practical engineering contexts.