对一般线性回归模型中有关参数估计分布的模拟问题，给出一种随机加权逼近的再构造方法。
A reconstructing method for random weighting approximations is proposed in approach to the distributions of the parameter estimates in general linear regression model.
把计算方法中的最小二乘法与泛函的抽象空间联系起来，得到最佳逼近问题的抽象提法，并在赋范线性空间探讨最小二乘法。
This article focuses on the topics that by connecting the the least-squares Minimization with the functional abstract space, the abstract definition of the best approximation problem is abtained.
本文先把问题转化为一个随机非线性规划问题，然后用逼近技术给出一个简单的求解方法。
We first formulate the problem as a stochastic nonlinear programming problem and then propose a new simple method for solving it via some approximation techniques.
应用推荐