Cross validation is a method to estimate the true error of a model by building models from subsets of training data and testing them on the remaining subsets. It provides a better estimate of how the model will generalize to new, unseen data compared to simply measuring error on the training data. Cross validation can also help evaluate which learning algorithm or parameter settings are best by comparing error rates from different options. Nested sub-processes in RapidMiner allow operators to contain additional processes that can be viewed by double clicking the icon.