案例教学法在“机器学习”课程中的应用与探索  

Application and Exploration of Case Teaching Method in the Course of Machine Learning

作　　者：王春玲[1] 沈法琳 韦颖 钱鹏[2] WANG Chunling;SHEN Falin(Anhui Sanlian University,Hefei 230601,China)

机构地区：[1]安徽三联学院现代康养产业学院,安徽合肥230601 [2]安徽三联学院,安徽合肥230601

出　　处：《通化师范学院学报》2025年第2期100-106,共7页Journal of Tonghua Normal University

基　　金：安徽省教育厅质量工程项目(2022zybj035,2023xxkc387,2023xjzlts091,2023jyxm0891,2023cxtd114,2023sdxx116);安徽三联学院质量工程项目(22zlgc108);教育部供需对接就业育人项目(2023122798976).

摘　　要：针对“机器学习”课程理论知识多、知识内容抽象、应用领域复杂、实践性较强等特点,设计了蘑菇识别系统案例.该案例先将多种传统的机器学习方法和基于卷积神经网络的深度学习方法在同一个任务上进行研究,然后采用UCI的Mushroom Data Set数据集,通过实验对比发现基于卷积神经网络构建的深度学习模型识别蘑菇是否有毒的准确率最高,最后选择最佳的卷积神经网络模型构建基于Tkinter的图形用户界面应用程序的蘑菇毒性识别系统.通过蘑菇识别系统研究,能够激发学生的学习兴趣和积极性,对比学院同一专业不同年级的学生成绩和学生竞赛成果,教学效果良好,可为“机器学习”课程实践教学和改革提供案例参考.knowledge content,complex application field and strong practicality,a case of mushroom recognition system is designed.This case first studies a variety of traditional machine learning methods and deep learning methods based on convolutional neural networks on the same task,and then uses UCI's Mushroom Data Set data set.Through experimental comparison,it is found that the deep learning model based on convolutional neural network has the highest accuracy in identifying whether mushrooms are toxic or not.Finally,the best convolutional neural network model is selected to construct a mushroom toxicity recognition system based on Tkinter's graphical user interface application.Through the research of mushroom recognition system,the students'interest and enthusiasm can be stimulated.Compared with the students'scores and students'competition results of the same major at different ages,it is found that the teaching effect is better,which provides a case reference for the practical teaching and reform of the course of Machine Learning.

关 键 词：机器学习 蘑菇识别 案例教学 

分 类 号：G642[文化科学—高等教育学]