论文部分内容阅读
Consider a classic regression Problem:Y =Xβ + □,(1.1)The regression coefficients β of (??) is often estimated with Ordinary LeastSquare(OLS) method:β=(XTX)-1XTy.(1.2)However,multi-coilinearity of X usually causes the variance of the coefficient estimation too large to get appropriate estimation and to achieve significance test of the coefficients.