论文部分内容阅读
最小二乘回归法在处理现场观测资料和模型实验数据中,是最佳的拟合方法之一。随着电子计算机技术的发展和对数据处理速度提高的要求,将最小二乘回归处理方法编译成通用计算程序,采用电子计算机进行数据处理,已成为分析数据资料所必需。笔者基于上述目的,应用最小二乘回归原理,分别用FOR T R A N语言编译了多元线性最小二乘回归法、二元线性有约束的最小二乘回归法、最小二乘逐步逼近法的电子计算机通用计算程序,并以波浪、水流共同作用下底部较细颗粒砂的起动实验数据处理为例,经计算应用结果良好。该计算程序可适用于多种观测数据和不同实验条件下各种因素组合的二元及多元函数的求解,其数据处理均可因不同的出发点借用不同的计算程序来完成。文中,除对计算的基本概念、计算疗法、计算步骤分别进行了讨论外,还对不同方法的统计检验结果进行了分析。三种计算方法的计算结果,是令人满意的。
The least square regression method is one of the best fitting methods in dealing with field observations and model experimental data. With the development of electronic computer technology and the request of increasing the data processing speed, it is necessary to compile the least square regression processing method into general computing program and use electronic computer to process data. Based on the above purpose, the author applies the principle of least-squares regression, compiles multiple linear least-squares regression method using FOR TRAN language, binary least-squares constrained least-squares regression method and least-squares stepwise approximation method, Program, and wave and water under the action of the bottom of the finer particles of sand start experimental data processing as an example, the application of the results are good. The calculation program can be applied to the solution of binary and multivariate functions of various observation data and combination of various factors under different experimental conditions. The data processing can be completed by borrowing different calculation programs from different starting points. In addition, the basic concepts, computational therapy, and calculation steps of the calculation are separately discussed, and the statistical test results of different methods are also analyzed. The calculation results of the three calculation methods are satisfactory.