A comparative study between a multi-stage adaptive pool testing model without test errors and the non-adaptive model

Abstract

Pool testing for presence or absence of a trait is less expensive, less time consuming and therefore more cost effective. This study presents a multi-stage adaptive pool testing estimator p ̂en of prevalence of a trait in the absence of test errors. Pool testing is more efficient, less expensive and less time consuming. An increase in the number of stages improves the efficiency of the estimator, hence construction of a multi-stage model. The study made use of the Maximum Likelihood Estimate (MLE) method and Martingale method to obtain the adaptive estimator and Cramer-Rao lower bound method to determine the variance of the constructed estimator. Matlab and R, statistical softwares were used for Monte-carlo simulation and verification of the model, then analysis and discussion of properties of the constructed estimator in comparison with the non-adaptive estimator in the literature of pool testing done alongside provision of the confidence interval of the estimator. Results have shown that as the number of stages increases, the efficiency of the multi-stage adaptive estimator in the absence of test errors also increases in comparison with the non-adaptive estimator in the absence of test errors. This makes the multi-stage adaptive estimator better than the corresponding non-adaptive estimator in the literature of pool testing.