step_distributed_lag — step_distributed

Generates distributed lag vectors. For regular spaced lags this uses an FFT based method which is faster and more memory efficient.

Usage

step_distributed_lag(
  .rec,
  terms,
  n_lag = 12L,
  lag_max = 86400L,
  knots = NA_real_,
  basis_matrix = NA_real_,
  intercept = FALSE,
  role = "predictor",
  ...
)

Arguments

.rec: the R6 recipe object.
terms: the unquoted names of the variables to use or a selector function. terms replaces the `...` of the recipes package but requires variables to be included within `c()`. For example to include variables x and y you would write `c(x,y)` in the hydrorecipes package.
n_lag: number of lags to calculate (enter either lag_max and n_lag or knots)
lag_max: maximum lag to calculate (enter either lag_max and n_lag or knots)
knots: specify the knot locations
basis_matrix: user specified basis_matrix
intercept: include intercept in basis matrix
role: character - the name of the role
...: additional arguments

Value

an updated recipe

References

Gasparrini, A., 2011. Distributed Lag Linear and Non-Linear Models in R: The Package dlnm. Journal of statistical software 465 43, 1–20.

Examples

formula <- as.formula(y~x)
rows <- 1e4

dat <- data.frame(x = rnorm(rows),
                  y = as.numeric(1:rows),
                  z = rnorm(rows))

frec = recipe(formula = formula, data = dat) |>
 step_distributed_lag(x, knots = hydrorecipes:::log_lags_arma(6, 800))