Exponential-type regression compromised imputation class of estimators
Ahmed Audu, Ran Vijay Singh
Abstract
Surveys such as medical, social science, etc conducted by human are often challenged by problem of non-response or missing observations. Different imputation strategies have been developed by several authors to take care of missing observation during analyses. Nevertheless, the estimators involved in some of these schemes depend on unknown parameters of the variable of interest which makes them impractical in real life. In this study, new classes of regression-type compromised imputation method which are free of unknown parameters have been presented. The properties (bias and MSE) of the class of estimators presented were derived up to first order approximation using Taylor series approach. Also, conditions for which the new estimators more efficient than other estimators considered in the study were also established. Numerical examples were conducted and the results revealed that the proposed class of estimators is more efficient.