Comparison between normal weights and conditional Bays weights in Iterative principal component estimators
Abstract: This paper discusses the problem of semi maulticollinearity in the nonlinear regression model (the multi-logistic regression model) When the dependent variable is a qualitative variable, the binary response is either equal to one for a response or zero for no response, Through the use of Iterative principal component estimatorsWhich are based on the normal weights and conditional Bays weights . If the appliede Estimates this model Through the use of two types of drugs concentrations thy concentration of ciprodar (variable X1) On a number of people with Patients with renal disease represent the dependent variable (The person heals from the disease , The person has not recovered from the disease )from through Mean Error Squares (MSE) The results were indicative of Iterative principal component estemaite Depending on the conditional Bays weights prefer the Iterative principal component estimators Depending on the the normal weights.