Stata 15 help for forecast estimates

[TS] forecast estimates -- Add estimation results to a forecast model

Syntax

Add estimation result currently in memory to model

forecast estimates name [, options]

name is the name of a stored estimation result; see [R] estimates store.

Add estimation result currently saved on disk to model

forecast estimates using filename [, number(#) options]

filename is an estimation results file created by estimates save; see [R] estimates save. If no file extension is specified, .ster is assumed.

options Description ------------------------------------------------------------------------- predict(p_options) call predict using p_options names(namelist[, replace]) use namelist for names of left-hand-side variables advise advise whether estimation results can be dropped from memory -------------------------------------------------------------------------

Description

forecast estimates adds estimation results to the forecast model currently in memory. You must first create a new model using forecast create before you can add estimation results with forecast estimates. After estimating the parameters of an equation or set of equations, you must use estimates store to store the estimation results in memory or use estimates save to save them on disk before adding them to the model.

Options

predict(p_options) specifies the predict options to use when predicting the dependent variables. For a single-equation estimation command, you simply specify the appropriate options to pass to predict. If multiple options are required, enclose them in quotation marks:

. forecast estimates ..., predict("pr outcome(#1)")

For a multiple-equation estimation command, you can either specify one set of options that will be applied to all equations or specify p options, where p is the number of endogenous variables being added. If multiple options are required for each equation, enclose each equation's options in quotes:

. forecast estimates ..., predict("pr eq(#1)" "pr eq(#2)")

If you do not specify the eq() option for any of the equations, forecast automatically includes it for you.

If you are adding results from a linear estimation command that forecast recognizes as one whose predictions can be calculated as x_t'b, do not specify the predict() option, because this will slow forecast's computation time substantially. Use the advise option to determine whether forecast needs to call predict.

If you do not specify any predict options, forecast uses the default type of prediction for the command whose results are being added.

names(namelist[, replace]) instructs forecast estimates to use namelist as the names of the left-hand-side variables in the estimation result being added. You must use this option if any of the left-hand-side variables contains time-series operators. By default, forecast estimates uses the names stored in the e(depvar) macro of the results being added.

forecast estimates creates a new variable in the dataset for each element of namelist. If a variable of the same name already exists in your dataset, forecast estimates exits with an error unless you specify the replace option, in which case existing variables are overwritten.

advise requests that forecast estimates report a message indicating whether the estimation results being added can be removed from memory. This option is useful if you expect your model to contain more than 300 sets of estimation results, the maximum number that Stata allows you to store in memory; see limits. This option also provides an indication of the speed with which the model can be solved: forecast executes much more slowly with estimation results that must remain in memory.

number(#), for use with forecast estimates using, specifies that the #th set of estimation results from filename be loaded. This assumes that multiple sets of estimation results have been saved in filename. The default is number(1). See [R] estimates save for more information on saving multiple sets of estimation results in a single file.

Examples

Setup . webuse klein2 . reg3 (c p L.p w) (i p L.p L.k) (wp y L.y yr) if year < 1939, endog(w p y) exog(t wg g) . estimates store klein

Create a new forecast model . forecast create kleinmodel

Add the stochastic equations fit using reg3 to the kleinmodel . forecast estimates klein

Add the advise to the above command to display special code to obtain fast predictions for the reg3 results . forecast estimates klein, advise


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index