摘要
文中利用矩阵的扫描运算,提出用主变量筛选法来对武器数据分析中出现的高维随机向量进行降维处理,并给出了一个算例。该方法是不同于主成分分析法的一种新的降维方法,它能有效地减小多重共线性问题带来的影响,尤其在处理数据多重相关性突出的武器费用数据时,该方法有着良好的效果,最后,作者用一个实用算例证明了其有效性和可行性。
In this paper, the calculation scheme of DR(dimension reduction) method in parameter optimization is discussed. By use the matrix scan, the author gave out a new method, the main variable filter method to lower the dimensions of the high dimensions variable in the analysis to the arm data. And the author also gave us a example, the method is a new method to lower the dimensions of data. Especially to the disposal of the arm expense data, the method can have good effect.
出处
《弹箭与制导学报》
CSCD
北大核心
2007年第1期220-222,共3页
Journal of Projectiles,Rockets,Missiles and Guidance
关键词
主变量筛选法
多重相关性
降维
贯用数据
main variable filter method
multiple relativity
dimension reduction
cost data