摘要
线性回归中,针对最小二乘法的两个替代准则—绝对离差和最小准则以及最大绝对离差最小准则,利用线性规划技术建立回归预测模型。实用分析表明,线性规划模型具有较好的预测效果,有郊地消除了统计数据中异常值对回归方程的影响。
This paper drafts to the least square method two substitution crlteria-total sum of square criterion and the moat greatly absolute dispersion minimum criterion, establishes regression forecast model by using linear programming. The example indicated that the linear programming model has the good forecast effect and eliminate the unusual statistical data influence on the regression equation effectively.
出处
《数学理论与应用》
2009年第2期64-67,共4页
Mathematical Theory and Applications
关键词
线性规划
最小二乘法
参数
线性回归
Linear programming
Least square method
Parameter
Linear regression
