Regression based thresholds in principal loading analysis

Principal loading analysis is a dimension reduction method that discards variables which have only small distorting effect on the covariance matrix. As special case, principal are not correlated with remaining ones. In multivariate linear regression other hand, predictors neither both nor dependent coefficients equal to zero. Hence, if goal select number of predictors, do correlate discarded as it also done in analysis. That methods same occurs for case zero correlation however. We contribute conditions under share variable selection. Further, we extend those provide choice threshold follows recommendations based simulation results so far.