基于“概念意象一多元表征一认知迁移”框架的贝叶斯公式教学设计

打开文本图片集

中图分类号：TP39；G434 文献标识码：A 文章编号：2096-4706（2026）04-0164-05

Teaching Design of Bayesian Formula Based on “Concept Image-Multiple Representation-Cognitive Transfer" Framework

WANG Shuaige （College of International Studies，National UniversityofDefense Technology，Nanjing21oo39，China）

Abstract： The Bayesian formula is characterized by high abstractness.Traditional teaching modes focus on symbolic derivation，whichmakesitdicultforsudentstoacieve“understanding-tranfer”inAtificialIntellgencesenarios.This paperconstructsathree-level progresive teaching frameworkof“conceptimage-multiplerepresentation-cognitive transfer” Thefirstlevelusesa“diseasescreening”magneticpuzleactivitytoactivate students’intuitivecognitionofprior probability likelihoodand posterior probability.Thesecond leveldevelops web interactiveanimations todynamicallydemonstrate the beliefupdating processin spam clasificationand reinfore the understanding of the Bayesian iterative mechanism.The third levelrelies onthe classic mushroom dataset to guide students to manuallybuilda Naive Bayes classifier and compare it with the classificationresultsof22-dimensional features implemented basedonscikit-learn，soastocompletethefar transfertohig dimensional real-world tasks.

Keywords： Bayesian formula; visualization; concept image; multiple representation; cognitive transfer

0 引言

贝叶斯公式作为概率论与数理统计课程的核心内容之一，因其高度抽象性与综合性，长期以来被学生视为“最难攻克的知识点”。（剩余5574字）