Resampling statistical techniques

With the availability of statistical softwares, resampling techniques which are computationally intensive are used more frequently. They are basically set of statistical inference methods based on new samples drawn from the initial sample or simulating datasets from an existing dataset, without any assumption on the underlying population.. These techniques include bootstrap and jackknife estimation, Monte Carlo permutation (randomization) tests and cross-validation method. Bootstrapping and jackknifing create 1000 or 5000 or more samples by drawing with replacement and they improve statistical precision of sample estimates (by computing bias and variance). Monte Carlo permutation (randomization) tests provide unbiased significance of the observed value of a test statistic by schuffling data without replacement. Cross validation involves repeatedly drawing samples from a training set of observations and examining the robustness and variability of a model on each sample in order to obtain additional insights.