CAT SWARM OPTIMIZATION FOR THE DETERMINATION OF STRATA BOUNDARIES

Abstract

A stratified random sampling method is preferred for selecting varied populations with outliers. As opposed to plain random sampling, stratified sampling increases statistical precision by reducing estimator variance. Before reducing the estimator's variance, stratum boundary identification and data apportionment must be solved. In this study, a Neyman allocation strategy is used to address the stratum boundary determination issue in mixed populations. In addition to evaluating CSO on two groups of people, a comparison study was conducted using Kozak, GA, PSO, and Delanius and Hodge's approaches. Compared to previous algorithms, the numerical results indicate that the proposed technique can select the best-stratified boundaries for various standard populations and test functions.