make_Q_splines_multi: Multivariate Spline Gibbs Model Pseudolikelihood Design...
In antiphon/PenGE: Penalised Gibbs Estimation

Description Usage Arguments Details Value

Construct the model matrix of multi-level multi-variate saturation model for penalised logistic regression pseudo-likelihood.

make_Q_splines_multi(x, knots1, knots2, spline_order = 3, sat1 = 1,
  sat2 = 1, sat1l = NULL, sat2l = NULL, auto_sat = TRUE,
  covariates, penalise_covariates = TRUE, rho_factor = 4,
  rho_N = NULL, dum_max = 10000, dum_min = 100, dummies = NULL,
  save_locations = TRUE, ..., verb = FALSE, cov_noise_sd = 0,
  fatal_col0 = FALSE, border_r = 0)

`x`	Point pattern data, preferrably a ppp-object. Say of p-types.
`knots1`	Intra-type knots, p vectors in a list. Include 0 and max R in each vector.
`knots2`	Inter-type knots, p*(p-1)/2 vectors in a list. Include 0 and max R in each vector.
`spline_order`	3 is cubic, 1 is tents. Default 3.
`sat1`	Intra-type saturations, vectors (for Saturation model)
`sat2`	Inter-type saturations, vectors (for Saturation model)
`sat1l`	More particular sat1, matrices (for Saturation model)
`sat2l`	More particular sat2, matrices (for Saturation model)
`auto_sat`	Determine saturations automatically?
`covariates`	list of im-objects to use as covariates
`penalise_covariates`	If FALSE, covariates are estimated without penalisation.
`rho_factor`	Dummy pattern intensity w.r.t. data
`rho_N`	Use this many dummies (vector pf length p)
`dum_max`	Upper limit for dummy count, per type (after rho_factor and rho_N)
`dum_min`	Lower limit for dummy count, per type (after rho_factor and rho_N)
`dummies`	Optional: (x,y,m) matrix of dummy locations to use.
`save_locations`	Store the locations in the output object? Used for CV
`...`	Omitted.
`verb`	Verbosity
`cov_noise_sd`	If > 0, add normal noise to covariates.
`fatal_col0`	Should we error on zero interaction columns (singularities).
`border_r`	Border correction range.