SIMULASI METODE RESAMPLING DAN PENDUGAAN DATA HILANG TERBAIK

Abstract

Data analysis has been very widely used in research, but problems with data such as small sample sizes and the presence of missing data cannot be avoided. The method of resampling and estimating missing data has advantages and disadvantages of each. Ratio estimator is an estimation method that is used when there are parameters that are difficult to know forecasts, therefore researchers want to know the best method of resampling and estimation of missing data by looking at the results of the ratio estimators and the various estimates of the ratio. This study uses a simulation method of resampling from data The Academic Performance Index (API) as a population using the R programming language. In this study, the Jackknife method produces the same estimated ratio with the Stratified Random Sampling method. The estimated ratios of the Jackknife and Bootstrap methods are relatively smaller than the Stratified Random Sampling method. In the Bootstrap method, the assumptions range depends on the number of repetitions, the more replications the assumptions range is relatively smaller. In the method of estimating missing data for this case, the Regression Imputation (Deterministic) method and the EM algorithm are better than the Regression Imputation (Stochastic) method due to the approximation of the approximation ratios and the relatively smaller assumptions of the ratio assumptions.