基于参数更新的灰色聚类概率模型评测研究生的创新能力
|
摘要
研究生的创新能力的评价,对于培养学校和研究生的职业发展都是非常重要的问题。本文认为专家的打分和灰色系统的评分系统的标准都存在概率误差的合理假设下,现在设二者都服从正态概率分布。然后,提出了基于概率的研究生创新能力的白化函数,以根据得分判别每人的创新能力的所属灰类。接着,在接受了新的信息以后,让打分系统基于Bayesian理论更新了评分系统的参数,即修正期望和方差,从而使得灰色系统的期望更准确,方差更小,这样能得出更准确的判别。在随后的两种评估模型的实例演示里,可以看到这个理论的优点,算法的高效性和稳健性。最后,进一步讨论了一种概率理论上的一致最优的截尾估计的优化估计方法。
It is very important to evaluate the innovative ability of postgraduate, to their cultivate university and career development of postgraduate themselves. In fact, both the mark of expert and grey clustering system exist errors and subjective factors, so it is reasonable to assume that expert's mark is normally probability distribution, so is the grey system. We propose a new probability model for grey clustering. Having new information, we renew the parameters of grey system based on Bayesian theory, that is to say, we have more accurate expect and smaller variance. So people may have more accurate judge method. Later,we give two examples,and we can easily see that our method is the latest, efficiently and robustly. At last, we propose a uniformly optimal truncating estimation on expect.
It is very important to evaluate the innovative ability of postgraduate, to their cultivate university and career development of postgraduate themselves. In fact, both the mark of expert and grey clustering system exist errors and subjective factors, so it is reasonable to assume that expert's mark is normally probability distribution, so is the grey system. We propose a new probability model for grey clustering. Having new information, we renew the parameters of grey system based on Bayesian theory, that is to say, we have more accurate expect and smaller variance. So people may have more accurate judge method. Later,we give two examples,and we can easily see that our method is the latest, efficiently and robustly. At last, we propose a uniformly optimal truncating estimation on expect.
引文
[1] Dickie C,Jay L.Innovation in postgraduate teaching:Mixed methods to enhance learning and learning about learning[J].Higher Education Research and Development,2010,29(1):29～43.
[2] Mudrak T,Wagenberg A,Wubben E.Assessing the innovative ability of FM teams:A review[J].Emerald Insight,2004,22:290～295.
[3] Jang S.ANFIS:Adaptive network-based fuzzy inference system[J].Sys.Man Cybern.IEEE Trans,1993,23:665～685.
[4] Abdul Shshed A,Longstaff A,Fletcher S.Thermal error model of machine tools based on ANFIS with fuzzy C-means clustering using a thermal imaging camera[J].Applied Mathematical Modelling,2015,39:1837～1852.
[5] 舒海翅,王新洲,花向红,田玉刚.模糊ISODATA中分类数C的确定[J].模糊系统与数学,2004,18:318～322.
[6] 邓聚龙.灰理论基础[M].武汉:华中科技大学出版社,2002.
[7] 唐晓清,谭明纯.简单序约束条件下参数估计的EM方法[J].湖南工程学院学报,2010,20(1):61～64.
[8] 王冰洁,唐晓清.根图的稳定性及其优化[J].西南师范大学学报(自科版),2017,42(4):14～19.
[9] 唐晓清,白延琴,刘念祖,刘莹.基于随机矩阵理论的 Markowitz组合投资模型[J].上海大学学报(自科版),2013,19(3):293～297.
[10] Tang X.New expected value expansions of rooted graphs[J].Acta Mathematicae Applicatae Sinica,English Series,2015,31(1):81～88.
[11] 唐晓清,王汉兴.识别药物最小有效剂量的逐步方法[J].药物评价研究,2010,33(1):58～62.
[12] 刘莹,唐晓清.图的双变量色多项式研究[J].湖南师范大学自然科学学报,2014,37(6):67～72.
[13] 唐晓清,刘念祖,王汉兴,白延琴.图的一类新双变量色多项式[J].兰州大学学报:自然科学版,2012,48(2):106～112.
[14] 唐晓清,刘念祖,王汉兴,白延琴.正则树的双变量色多项式研究[J].应用数学学报,2013,36(4):761～768.
[15] 刘莹,唐晓清.含根网络的可靠性与稳定性研究[J].重庆师范大学学报(自科版),2017,34(5):65～71.
[16] 刘莹,唐晓清.正则q-树根图的可靠性研究[J].东北师大学报(自科版),2014,47(1):17～21.
[17] 刘念祖,唐晓清,王汉兴.正则q-树的双概率可靠性研究[J].西南师范大学学报(自科版),2013,38(12):24～27.
[18] 茆诗松,王静龙,濮晓龙.高等数理统计[M].北京:高等教育出版社,2006.
[19] 刘莹,唐晓清,王双成.基于特征方程的某些图双概率可靠性的统一形式[J].西南师范大学学报(自科版),2018,43(10):18～21.
[2] Mudrak T,Wagenberg A,Wubben E.Assessing the innovative ability of FM teams:A review[J].Emerald Insight,2004,22:290～295.
[3] Jang S.ANFIS:Adaptive network-based fuzzy inference system[J].Sys.Man Cybern.IEEE Trans,1993,23:665～685.
[4] Abdul Shshed A,Longstaff A,Fletcher S.Thermal error model of machine tools based on ANFIS with fuzzy C-means clustering using a thermal imaging camera[J].Applied Mathematical Modelling,2015,39:1837～1852.
[5] 舒海翅,王新洲,花向红,田玉刚.模糊ISODATA中分类数C的确定[J].模糊系统与数学,2004,18:318～322.
[6] 邓聚龙.灰理论基础[M].武汉:华中科技大学出版社,2002.
[7] 唐晓清,谭明纯.简单序约束条件下参数估计的EM方法[J].湖南工程学院学报,2010,20(1):61～64.
[8] 王冰洁,唐晓清.根图的稳定性及其优化[J].西南师范大学学报(自科版),2017,42(4):14～19.
[9] 唐晓清,白延琴,刘念祖,刘莹.基于随机矩阵理论的 Markowitz组合投资模型[J].上海大学学报(自科版),2013,19(3):293～297.
[10] Tang X.New expected value expansions of rooted graphs[J].Acta Mathematicae Applicatae Sinica,English Series,2015,31(1):81～88.
[11] 唐晓清,王汉兴.识别药物最小有效剂量的逐步方法[J].药物评价研究,2010,33(1):58～62.
[12] 刘莹,唐晓清.图的双变量色多项式研究[J].湖南师范大学自然科学学报,2014,37(6):67～72.
[13] 唐晓清,刘念祖,王汉兴,白延琴.图的一类新双变量色多项式[J].兰州大学学报:自然科学版,2012,48(2):106～112.
[14] 唐晓清,刘念祖,王汉兴,白延琴.正则树的双变量色多项式研究[J].应用数学学报,2013,36(4):761～768.
[15] 刘莹,唐晓清.含根网络的可靠性与稳定性研究[J].重庆师范大学学报(自科版),2017,34(5):65～71.
[16] 刘莹,唐晓清.正则q-树根图的可靠性研究[J].东北师大学报(自科版),2014,47(1):17～21.
[17] 刘念祖,唐晓清,王汉兴.正则q-树的双概率可靠性研究[J].西南师范大学学报(自科版),2013,38(12):24～27.
[18] 茆诗松,王静龙,濮晓龙.高等数理统计[M].北京:高等教育出版社,2006.
[19] 刘莹,唐晓清,王双成.基于特征方程的某些图双概率可靠性的统一形式[J].西南师范大学学报(自科版),2018,43(10):18～21.