基于RVM机器学习方法的高校理论与实验教学质量评价模型  

Universities theoretical and experimental teaching quality evaluation model based on RVM machine learning method

作　　者：傅翠霞 罗亦泳 FU Cuixia;LUO Yiyong(College of foreign Languages,East China University of Technology,Nanchang 330013,China;School of Geodesy and Geomatics,Wuhan University,Wuhan 430079,China)

机构地区：[1]东华理工大学外国语学院,江西南昌330013 [2]武汉大学测绘学院,湖北武汉430079

出　　处：《现代电子技术》2019年第13期181-186,共6页Modern Electronics Technique

基　　金：国家自然科学基金项目(41204003);江西省教育科学“十二五”规划课题(13YB126);江西省高等学校教学改革研究课题(JXJG-16-6-28,JXJG-16-6-25)~~

摘　　要：为提高高校理论教学与实验教学质量评价的精度及可靠性,基于机器学习新算法RVM构建高校理论教学与实验教学质量评价新模型。对RVM的核函数及参数确定方法进行改进,利用遗传算法优化的MK-RVM模型分别建立高校理论与实验教学质量评价的MK-RVM新模型,并构建基于置信区间的结果可靠性分析方法。实验结果表明,高校理论教学与实验教学质量评价MK-RVM模型的各项精度指标均表现优秀,精度等级达到“好”;训练集及测试集的教学质量评价真实值均在MK-RVM估计的95%置信度置信区间内,证实基于MK-RVM的教学质量评价结果具有较好的可靠性。因此,RVM算法为高校理论教学与实验教学质量评价提供了一种新的有效途径。The new quality evaluation models of theoretical teaching and experimental teaching in universities are built on the basis of relevance vector machine(RVM) to improve the accuracy and reliability of the quality evaluation of theoretical teaching and experimental teaching in universities. The kernel function and parameter determination method of RVM are improved,and the MK - RVM model optimized by genetic algorithm is established. The new MK - RVM models for quality evaluation of theoretical teaching and experimental teaching in universities are set up respectively,and the result reliability analysis method based on confidence interval is constructed. The experimental results show that all the precision indicators of MK- RVM model based quality evaluation of theoretical teaching and experimental teaching in universities are excellent,and the precision level can reach up to "good". The real values of teaching quality evaluation of training set and test set are within the 95% confidence interval of MK-RVM estimation,which confirms that the MK-RVM based evaluation result of teaching quality has high reliability. Thus,the RVM algorithm provides a new effective way for the quality evaluation of theoretical teaching and experimental teaching in universities.

关 键 词：相关向量机(RVM) 教学质量评价 精度分析 可靠性分析 遗传算法 置信区间 

分 类 号：TN911-34[电子电信—通信与信息系统]