predictBN: Downscale with Bayesian Networks

In MNLR/BNdownscaleR: Statistical Downscaling with Bayesian Networks.

Description

Downscale with Bayesian Networks

Usage

predictBN(cbn, x, y = NULL, output = "probabilities",
  prediction.type = "exact", event = "1", threshold.vector = NULL,
  output.attr.evidence = FALSE, cl = NULL, stop.cluster = TRUE,
  parallelize = FALSE, n.cores = NULL, cluster.type = "FORK")

Arguments

cbn	Climatic Bayesian Network, as returned by either buildCBN() or buildDynamicCBN().
x	may be output of prepareNewData for predictive non-past, NULL for a no-G dynamic CBN or a single dataset for a G dynamic CBN (i.e. dynamic CBN do not accept members).
output	Options are: "probabilities", "event" and "probabilities.list". You should probably not use the last one. Note that if prediction.type = "simulation", output is forced to "event".
prediction.type	Options are "exact", "approximate" and "simulation". Exact inference requires a compilable junction.

Author(s)

M.N. Legasa

MNLR/BNdownscaleR documentation built on June 12, 2019, 1:58 p.m.