有调整因子的回归弱相关估计

针对回归分析中自变量数据间的共线性问题，本文基于降低解释变量间相关程度的弱相关思想，构造出一个新的回归估计式。并证明了估计式与最小二乘相比，具有均方差意义下 的优良性。还证明了相对于最小二乘估计，新估计式是一种模压缩。该估计式具有意义直观明了，便于分析，应用的特性。

REDUCED CORRELATION REGRESSION WITH REGULATING FACTOR

Based on the notion to lessen the multicollinearity effect in regression analysis by reducing the correlation between variables to reduced correlation,this paper gives a new regression estimator which proves to be superior to the least square estimator in mean square error(MSE)efficiency,The paper also proves that,unlike the leaest square estimator,the newestimator turns out to be a module compression.