摘要
通过引入支持向量回归(Support Vector Regression,SVR)的大数据研究方法,使用Gaussian核搭建模型为满意度预测提出一种方案,分析M大学6年研究生教育满意度调查数据发现:导师职业水平、指导方向与频率、专业教学、学术支撑、教学科研软件支撑、管理服务等是影响研究生教育满意度的主要因素,而专业教学与学术支撑对不同主体和层面的满意度均有影响,从学生背景的主成份降维投影和样本点耦合欧氏距离中发现每类样本明显被求学意愿所聚类,导师培养是研究生教育满意度较为独立且非常重要的影响因素。因此,导师需根据不同学习动因对学生进行分类指导、校内不同层面机构和部门开展工作需有所侧重,以提高研究生教育满意度。
By introducing the big data research method of Support Vector Regression(SVR)and using the Gaussian kernel building model to propose a scheme for satisfaction prediction,the data of the 6-year graduate student education satisfaction survey of the M University was analyzed.It is found that the supervisor’s professional level,the study direction and guidance frequency,the teaching quality,the academic support,the teaching and research software facilities,and the management service are the main factors influencing the satisfaction of graduate education.However,the teaching quality and academic support have influence on the satisfaction at all levels.It is also found that each category of samples is obviously clustered by the willingness to study,and the tutor training is a relatively independent and very important influencing factor of the satisfaction of graduate education.Therefore,instructors need to categorize and guide students according to different learning motivations,and different levels of institutions and departments within the university need to focus their work in order to improve the satisfaction of graduate education.
作者
杨珪
戴睿
高堃原
贺喆南
何力
张鲜元
崔亚娟
陈钊宇
YANG Gui;DAI Rui;GAO Kunyuan;HE Zhenan;HE Li;ZHANG Xianyuan;CUI Yajuan;CHEN Zhaoyu(Department of Development planning,Sichuan University,Chengdu,Sichuan 610065,China;School of Computer Science,Sichuan University,Chengdu,Sichuan 610065,China;Department of Philosophy,Sichuan University,Chengdu,Sichuan 610065,China)
出处
《宜宾学院学报》
2021年第10期95-103,共9页
Journal of Yibin University
关键词
研究生
满意度
调查
大数据分析方法
SVR
graduate students
satisfaction
survey
big data analysis method
SVR