摘要
论证了实现线性回归最优预测的数学前提是自变量与因变量都服从同一分布的原理,推出了应用二步正态转换法实现这一原理的回归预预测法,通过二道典型数据预测结果的对比,体现了新方法的稳健性。
This paper has proved the principle that the mathematical precondition for linear regression to make optimum forecast is that dependent and independent variables must be subordinated to the.same distribution.The regression forecast method to realize the mathematical precondition with twostep normal transformation is produced.By comparing the predicted results from two set of typical data, the new method is proved to be robust.
出处
《物探化探计算技术》
CAS
CSCD
1991年第4期319-329,共11页
Computing Techniques For Geophysical and Geochemical Exploration