基于数据拟合的有确定度逆向云算法  

Data fitting based backward cloud algorithm with certainty degree

作　　者：陈昊[1] 李兵[2] Chen Hao Li Bing(School of Computer Science ＆ Information Engineering, Hubei University, Wuhan 430062, China International School of Software Wuhan University, Wuhan 430072 , China)

机构地区：[1]湖北大学计算机与信息工程学院,武汉430062 [2]武汉大学国际软件学院,武汉430072

出　　处：《计算机应用研究》2016年第12期3663-3666,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(61273216);湖北省教育厅科学技术研究重点资助项目(D20141005)

摘　　要：逆向云算法实现了定量数据到定性概念的转换,对云模型的研究有着重要的理论意义和应用价值。现有的有确定度逆向云算法采用均值法求期望Ex的估计值,当云滴样本数较少时将会出现较大误差。提出了一种通过非线性最小二乘法数据拟合高斯云的期望曲线求得期望Ex的估计值方法,再通过高斯分布参数估计法得到熵En和超熵He的估计值。通过误差比较实验,验证了改进的算法能减小参数估计误差,提高算法精度。Backward cloud algorithm can effectively transform quantitative values into qualitative concepts, which is significant to the theoretical research and practical applications of cloud model. The existing backward cloud algorithm with certainty degree employs the method of mean value to find the estimate of expected value （ Ex）, and large error will appear when the number of droplet samples is small. This paper proposed an approach that the estimate of Ex was found using an expectation curve of non-linear least squares data fitting Gauss cloud, and entropy （En） and hyper entropy （He） could be further found with Gaussian distribution parameters estimation. After the comparison of errors, it proves that the developed algorithm can reduce the error of parameter estimation and improve the algorithm precision.

关 键 词：云模型 高斯云 逆向云 数据拟合 误差分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]