A new shrinkage method for higher dimensions regression model to remedy of multicollinearity problem

This research seeks to present new method of shrinking variables select some basic from large data sets. shrinkage estimator is a modification (Ridge and Adaptive Lasso) regression in the presence mixing parameter that was calculated Elastic-Net. The Proposed called (Improved Mixed Shrinkage Estimator (IMSHE)) handle problem multicollinearity. In practice, it difficult achieve required accuracy efficiency when dealing with big set, especially case multicollinearity between explanatory variables. By using Basic methods (Lasso, Lasso Elastic Net) comparing their results New (IMSH) applied set obesity -related containing (52) for sample (112) observations. All have also been compared through Mean Square Error (MSE) criterion Cross Validation Parameter (CVP). showed best among four Lasso, Net IMSH) IMSH method, as corresponds lowest based on cross-validation test proposed achieved optimal (λ = 0.6932827) according (CVP) test, leads minimum value mean square error equal (0.2576002). regularization increases, decreases become zero, so ideal number after (p=6).