A moderated arrow puts the grid's own values inside \(B\), so there is no one covariance
to draw from. Each year is drawn instead in the order series_order gives, every cell
from what points into it and an innovation, which is the model the density evaluates. Years
already known are held as is.
Usage
draw_dsem_recursive(
dsem_model,
arrow_value,
mu_grid,
n_sims,
n_cond = 0,
x_known = NULL
)
Arguments
- dsem_model
Output of read_dsem_arrows.
- arrow_value
Values of the arrows, from get_dsem_arrow_values.
- mu_grid
Matrix [year, series] of series means.
- n_sims
Number of replicates.
- n_cond
Leading years kept at x_known rather than drawn.
- x_known
Matrix [year, series] read for those years.
Value
Array [year, series, sim].