针对输入、输出变量及回归系数都是LR型模糊数的模糊线性回归模型的参数估计,提出了一种新方法,就是通过将模糊观测数据用区间来表示,然后利用区间的左、右端点和中点的数据集求出传统线性回归模型相应的回归系数,从而得到模糊线性回归模型的最小二乘估计,当输入、输出变量及回归系数都退化为清晰数时,该估计就是传统的最小二乘估计,并通过对实例,说明了该模型应用方便,而且拟舍效果较好。
For parameter estimation of fuzzy linear regression model with input, output variables and regression coefficients are LR typed fuzzy numbers, a new method is proposed in the paper. By using intervals to represent the observed fuzzy data, then use the left, right point and the midpoint data sets of the interval to derive corresponding regression coefficients of conventional linear regression model. So that the least squares estimations of fuzzy linear regressive model are obtained. When the input, output variables and regression coefficients are reduced to clear numbers, the estimation is the traditional least squares estimation. An example is used to illustrate that the application of the model is easily and fitting results are good.