In forecasting, the process of testing how accurate a model is for making forecasts. The sample data are often split into two segments, one used to estimate the parameters of the model, and the other, the holdout data, used to test the forecasts made with the model. The many variations of validation include cross-validation, n-way validation, and the jackknife. Validation can also be used to assess the usefulness of the parameters of a forecasting model. Armstrong (2001d) describes various approaches to validation.