This section describes procedures for forecasting and computing fitted values from a single equation. The techniques described here are for forecasting with equation objects estimated using regression methods. Forecasts from equations estimated by specialized techniques, such as ARCH, binary, ordered, tobit, and count methods, are discussed in the corresponding sections.

Forecasting from a series using exponential smoothing methods is explained in “Exponential Smoothing” , and forecasting using multiple equations and models is described in “Models”. Additional tools for performing forecast evaluation are described in “Forecast Evaluation”