L function

Lag function for autoregressive models

Creates lagged variables for use with bayesGAM, including the functionality to create lags for each specified subject if desired. The input data must be pre- sorted according by time, and within each subject id if specified.

L(x, k = 1, id = NULL)

Arguments

• x: numeric vector
• k: integer vector of lagged variables to create
• id: optional identification number for each subject

Returns

numeric vector or matrix of the lagged variable(s)

Examples

x <- rnorm(20)
id <- rep(1:4, each=5)
L(x, 1:2, id)

# autoregressive
ar.ols(lh, demean = FALSE, intercept=TRUE, order=1)
f <- bayesGAM(lh ~ L(lh), family=gaussian)
coef(f)

References

Zeileis A (2019). dynlm: Dynamic Linear Regression. R package version 0.3-6