One simple yet effective method for testing estimation sensitivity is to run a bootstrap analysis. This is when a sample or set of samples is intentionally removed from a dataset, an estimation is performed in the location of the missing samples, and the estimated result is compared back to the samples.
The tools and workflow built into Leapfrog EDGE make this a simple operation to perform at a block level by comparing the results of two different estimations.
1. Create a standard estimator
The first step to running a bootstrap is to create an estimator using all data required. For this example I will use the LMS1 domain from the Leda training dataset, and will be estimating the Zn grades within that volume.