predict.ssnbayes {SSNbayes}

Performs spatio-temporal prediction in R using an ssnbayes object from a fitted model.

Description

It will take an observed and a prediction data frame. It requires the same number of observation/locations per day. It requires location id (locID) and points id (pid). The locID are unique for each site. The pid is unique for each observation. Missing values are allowed in the response but not in the covariates.

Usage

## S3 method for class 'ssnbayes'
predict(
  object = object,
  ...,
  path = path,
  obs_data = obs_data,
  pred_data = pred_data,
  net = net,
  nsamples = nsamples,
  addfunccol = addfunccol,
  locID_pred = locID_pred,
  chunk_size = chunk_size,
  seed = seed
)

Arguments

object

A stanfit object returned from ssnbayes

...

Other parameters

path

Path with the name of the SpatialStreamNetwork object

obs_data

The observed data frame

pred_data

The predicted data frame

net

(optional) Network from the SSN object

nsamples

The number of samples to draw from the posterior distributions. (nsamples <= iter)

addfunccol

The variable used for spatial weights

locID_pred

(optional) the location id for the predictions. Used when the number of pred locations is large.

chunk_size

(optional) the number of locID to make prediction from

seed

(optional) A seed for reproducibility

Details

The returned data frame is melted to produce a long dataset. See examples.

Value

A data frame with the location (locID), time point (date), plus the MCMC draws from the posterior from 1 to the number of iterations. The locID0 column is an internal consecutive location ID (locID) produced in the predictions, starting at max(locID(observed data)) + 1. It is used internally in the way predictions are made in chunks.

Author(s)

Edgar Santos-Fernandez

Examples

#require('SSNdata')
#clear_preds <- readRDS(system.file("extdata/clear_preds.RDS", package = "SSNdata"))
#clear_preds$y <- NA
#pred <- predict(object = fit_ar,
#                 path = path,
#                 obs_data = clear,
#                 pred_data = clear_preds,
#                 net = 2,
#                 nsamples = 100, # numb of samples from the posterior
#                 addfunccol = 'afvArea', # var for spatial weights
#                 locID_pred = locID_pred,
#                 chunk_size = 60)