Post-MCMC per-actor time-varying slopes

Computes a smoothed per-actor time-varying slope coefficient on a slice of the combined dyadic design array (fit$Xlist). For kind = "row", each row-actor \(i\) gets a length-\(T\) slope path \(\beta_{i,t}\) on the across-column mean of design slice covariate_idx, fit by ridge-penalised least squares on the residual after the main MCMC linear predictor; kind = "col" uses the across-row mean for each column-actor.

Usage

per_actor_slopes(fit, kind = c("row", "col"), covariate_idx = 1L, lambda = 1)

Arguments

fit

A fitted lame object.

kind

"row" (default; per-row-actor slopes) or "col" (per-column-actor slopes).

covariate_idx

Integer index into the third (covariate) dimension of the combined dyadic design array fit$Xlist. Slice 1 is the intercept; dyadic covariates follow. Default 1L.

lambda

Non-negative smoothing parameter (first-difference penalty across periods). Default 1.

Value

A list with slopes (an \(n_{actors} \times T\) matrix), kind, covariate_idx, lambda, and label. Class "per_actor_slopes".

Examples

# \donttest{
data(YX_bin_list)
fit <- lame(YX_bin_list$Y, YX_bin_list$X, family = "binary", R = 0,
            nscan = 100, burn = 25, odens = 5, verbose = FALSE)
#> Warning: `family` = "binary" but `Y` contains values other than 0/1.
#> ℹ `Y` will be thresholded to `1 * (Y > 0)`; if you meant counts, use "poisson",
#>   or "ordinal"/"normal" as appropriate.
# post-MCMC per-row-actor slopes on the first dyadic covariate
pas <- per_actor_slopes(fit, kind = "row", lambda = 1)
dim(pas$slopes)
#> [1] 50  4
# }