Title: Spatiotemporal Propagation for Multivariate Bayesian Dynamic Learning
Version: 0.0-2
Description: Implementation of the Forward Filtering Backward Sampling (FFBS) algorithm with Dynamic Bayesian Predictive Stacking (DYNBPS) integration for multivariate spatiotemporal models, as introduced in "Adaptive Markovian Spatiotemporal Transfer Learning in Multivariate Bayesian Modeling" (Presicce and Banerjee, 2026+) <doi:10.48550/arXiv.2602.08544>. This methodology enables efficient Bayesian multivariate spatiotemporal modeling, utilizing dynamic predictive stacking to improve inference across multivariate time series of spatial datasets. The core functions leverage 'C++' for high-performance computation, making the framework well-suited for large-scale spatiotemporal data analysis in parallel computing environments.
LinkingTo: Rcpp, RcppArmadillo
Imports: spBPS, Rcpp (≥ 1.1.1), foreach, tictoc, abind
Suggests: doParallel, mniw, MBA, ggplot2, patchwork, reshape2, knitr, rmarkdown
License: GPL (≥ 3)
Encoding: UTF-8
RoxygenNote: 7.3.3
VignetteBuilder: knitr
URL: https://lucapresicce.github.io/spFFBS/
NeedsCompilation: yes
Packaged: 2026-04-22 11:13:49 UTC; presi
Author: Luca Presicce ORCID iD [aut, cre]
Maintainer: Luca Presicce <l.presicce@campus.unimib.it>
Repository: CRAN
Date/Publication: 2026-04-22 13:50:02 UTC

spFFBS: Spatiotemporal Propagation for Multivariate Bayesian Dynamic Learning

Description

logo

Implementation of the Forward Filtering Backward Sampling (FFBS) algorithm with Dynamic Bayesian Predictive Stacking (DYNBPS) integration for multivariate spatiotemporal models, as introduced in "Adaptive Markovian Spatiotemporal Transfer Learning in Multivariate Bayesian Modeling" (Presicce and Banerjee, 2026+) doi:10.48550/arXiv.2602.08544. This methodology enables efficient Bayesian multivariate spatiotemporal modeling, utilizing dynamic predictive stacking to improve inference across multivariate time series of spatial datasets. The core functions leverage 'C++' for high-performance computation, making the framework well-suited for large-scale spatiotemporal data analysis in parallel computing environments.

Author(s)

Maintainer: Luca Presicce l.presicce@campus.unimib.it (ORCID)

See Also

Useful links:

compute_Wt: Dynamic Bayesian Predictive Stacking Weights

Description

Computes predictive stacking weights from the forward-filter objects. Provides a simplified and user-friendly interface with optional parallelization.

Usage

compute_Wt(out_FF, tau, phi, parallel = FALSE, verbose = TRUE)

Arguments

out_FF

Output of spFF3 (list of filtering results).

tau

Vector of tau grid values (only used for column naming).

phi

Vector of phi grid values (only used for column naming).

parallel

Logical; use parallel backend (foreach + doParallel)? Default FALSE.

verbose

Logical; print progress messages to the console? Default TRUE.

Value

Matrix of weights of size n x J

Function to sample integers (index)

Description

Function to sample integers (index)

Usage

sample_index(size, length, p)

Arguments

size

integer dimension of the set to sample

length

integer number of elements to sample

p

vector sampling probabilities

Value

vector sample of integers

spFFBS: Spatiotemporal Bayesian Pipeline (friendly interface)

Description

A user-friendly, modular wrapper for running a Bayesian spatiotemporal filtering + weighting pipeline, with optional backward sampling, forecasting, and spatial interpolation.

Usage

spFFBS(
  Y,
  G,
  P,
  D,
  grid = list(tau = NULL, phi = NULL),
  prior,
  do_BS = FALSE,
  do_forecast = FALSE,
  do_spatial = FALSE,
  L = 200,
  tnew = NULL,
  spatial = NULL,
  num_threads = 1,
  verbose = TRUE
)

Arguments

Y

Response data (3D array or cube).

G

System matrix (cube).

P

Observation matrix (cube).

D

Spatial distance matrix.

grid

List with elements:

  • tau: numeric vector

  • phi: numeric vector

prior

Prior list for forward filter (m, C, nu, Psi)

do_BS

Logical: run backward sampling? (default: FALSE)

do_forecast

Logical: run temporal forecasts? (default: FALSE)

do_spatial

Logical: run spatial interpolation? (default: FALSE)

L

Number of posterior samples (default 200)

tnew

Forecast horizon (default 5)

spatial

Optional list for spatial: list(crd = , crdtilde = , Xtilde = )

num_threads

Number of cores for parallel computing (default: 1)

verbose

Logical; print progress messages to the console? (default: TRUE)

Value

A list with the components executed according to the flags.

Examples

n <- 50
t <- 5
p <- 2
q <- 2

Y <- array(rnorm(n*q*t), dim = c(n, q, t))
P <- array(rnorm(n*(p+n)*t), dim = c(n, (p+n), t))
G <- array(rnorm((p+n)*(p+n)*t), dim = c((p+n), (p+n), t))
coords <- matrix(runif(n*2), ncol = 2)
D <- as.matrix(dist(coords))

priors <- list("m" = matrix(0, n+p, q),
               "C" = diag(p+n),
               "nu" = 3,
               "Psi" = diag(q))

hyperpar <- list(tau = 0.5, phi = 1)

res <- spFFBS(Y = Y, G = G, P = P, D = D, grid = hyperpar, prior = priors)