曲线拟合的逆向云改进算法

An improved algorithm of backward cloud based on curve fitting

逆向云发生器是从定量数据到定性概念的转化模型。在逆向云生成算法中,Ex的估计值对En和He的估计有比较大的影响,直接使用样本均值进行参数估计会导致较大的误差。本文通过深入研究云模型的相关算法,对逆向云生成算法进行改进。文中利用样本均值作为非线性拟合函数的初值,把拟合函数的返回参数作为Ex的估计值,改善了随机选取拟合初值造成结果不稳定现象,实现对逆向云参数优化。通过与其他几种逆向云生成算法进行对比,仿真结果表明:改进后的逆向云生成算法有较好的稳定性并且有较高的精度。

Backward cloud generator is a model,which transforms quantitative values into qualitative concepts. The estimated value of Exhas a great impact on Enand Hein backward cloud generated algorithm. The direct use of the sample mean will lead to large errors. This paper makes the more thorough research to the algorithm of cloud model.Furthermore,the backward cloud generated algorithm has been improved. The sample mean is used as the initial value of the nonlinear fitting function. Then,the fitting function returns parameter as estimated value of Ex. This can improve the instability of randomly selected the fitting initial value and achieve parameters optimization of the backward cloud algorithm. Compared with other several backward cloud generation algorithms,simulation result shows that the improved algorithm has better stability and higher accuracy.