面向自动化的机器学习实验平台设计  

Machine learning experimental platform design for automation

作　　者：杜晓冬[1] 王俊[1] DU Xiao-dong;WANG Jun(Department of Basic Teaching and Experiment Center,Hefei University, Hefei 230601, China)

机构地区：[1]合肥学院基础教学与实训中心,安徽合肥230601

出　　处：《湖北师范大学学报（自然科学版）》2021年第2期67-75,共9页Journal of Hubei Normal University：Natural Science

基　　金：国家自然科学基金面上项目(61973295);合肥学院质量工程项目(2019hfjyxm40)。

摘　　要：近年来,随着人工智能的发展及其在各领域中的广泛应用,机器学习技术逐渐成为高校计算机科学相关专业的教学重点。在自动化控制中,机器学习技术可以有效提升控制水平,因此融合传统的自动化技术和先进的机器学习技术可推进自动化的发展。为在自动化学科中辅助开展机器学习的实验教学,设计了一种基于MATLAB的机器学习实验平台,专门面向无机器学习研究基础的学习者。该平台界面友好,操作简单,具有直观的图像展示功能,可使学生初步了解和掌握支持向量机、K最近邻、决策树和朴素贝叶斯四种分类模型,并理解这些分类器的特性、优缺点与应用领域,同时培养学生对机器学习技术的学习兴趣。In recent years,with the development of artificial intelligence and its wide application in various fields,machine learning technology has gradually become the focus of teaching for the computer science and other related majors in universities.Since the utilization of machine learning technologies can improve the effect of automation,combining machine learning can promote the development of automation.In order to assist the experimental teaching of machine learning technology in the automation discipline,a Matlab-based machine learning experimental platform is developed,which is specialized for learners without machine learning research foundation.It has a friendly interface and it is simple to operate.Through the visualized image display,students can have a preliminary understanding of the four classification models including the support vector machine,K-nearest neighbors,decision tree and naive bayes.With the aid of the developed platform,students can understand the characteristics,advantages and disadvantages,applications of these classifiers.At the meantime,students’interest in machine learning is also cultivated.

关 键 词：自动化 人工智能 机器学习 实验教学 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]