bt.interval: Bootstapping Predictive Intervals

In google-trends-v1/gtm: Functions for Economic Nowcasting with Google Trends Data

Description

This function constructs 95% predictive intervals using bootstrapping. The bootstrapping process is done using block bootstrapping, in order to account for serial correlation in the time series data. The predictive intervals then can be visualized with plot.BtInterval function.

Usage

bt.interval(.data = NULL, boot = 100, forecast = "model")

Arguments

.data	an object of class "Maeforecast" returned from maeforecast.
boot	number of bootstrapped versions to generate. This cannot be too small in order for the predictive intervals to be meaningful. Default is set to be 100.
forecast	whether the original point forecasts ("model") should be reported or a new set of forecasts generated by bootstrapping aggregation ("mean"). Default is "model".

Details

This function automatically extracts the model information from the argument .data, including model type, forecasting window, window size, forecasting horizon, and so on. Users will only have to indicate the desirable number of bootstrapped versions to generate, and the function will take care of the rest.

Value

This function returns an object of class "BtInterval" that contains the following components

Intervals	a 3-column matrix that contains the lower bound, point forecasts and the upper bound.
Data	the original data matrix used in the model.
Model	the original model information.
Bootstrapped	all bootstrapped forecasts

Author(s)

Zehua Wu

See Also

plot.BtInterval

google-trends-v1/gtm documentation built on June 5, 2019, 5:13 p.m.