The primary objectives of this research are to develop andstudy estimators for generalized linear measurement errormodels when the mean function contains error-free predictorsas well as predictors measured with error and interactions between error-free and error-prone predictors. Attention is...

Most variable selection techniques focus on first-order linear regression models. Often, interaction and quadratic terms are also of interest, but the number of candidate predictors grows very fast with the number of original predictors, making variable selection more difficult. Forward...