Research And Application Of Synthesized Ridge Estimation With Compression Coefficient

Linear regression model in economy,medicine,management,engineering and other fields have important applications,and in the field of modern mathematical statistics plays an irreplaceable role.But when they return when there are multicollinearity issues among the independent variables,least square estimation of performance becomes very poor,all kinds of biased estimators have been developing rapidly.Regression linear regression model between independent variables the multicollinearity problem,first of all,in synthesis range estimate research foundation,introduces is similar in the Stein compression estimate the compressibility coefficient improves the comprehensive range to estimate,obtained the h-D synthesis range estimate and the related nature,not only perfect in solving multicollinearity problem,and proved that in mean square error sense it is better than least squares estimation and synthesized ridge estimate.Secondly,applies its thought in having in the homogeneous equality constraint return model,proposed the h-D synthesis condition range estimated,and gives many estimates the nature to explain its rationality.Finally,On the basis of the h-D comprehensive ridge estimation,combined with the features of the data model,to the h-D synthesis range estimated the data deletion model the strong influence question conducts the research.h-D synthesized ridge estimate is proved by examples than least squares estimation and synthesized ridge estimate,roves h-D synthesized Ridge estimate error smaller,more practical significance.