predict.spacious: Make predictions with spacious fit
In jarad/spacious: Block composite spatial analysis.

Description Usage Arguments See Also

Predicts values at newS locations using a spacious fit.

## S3 method for class 'spacious'
predict(object, newdata, newS, newB, D,
        opts = list(type="block"),
        interval = "none", level = 0.95,
        nthreads = 1, gpu=FALSE,
        engine, ...)

`object`	model fit with `spacious`.
`newdata`	data frame of variables used in the model fit. Not required with intercept only models.
`newS`	an n by 2 matrix of spatial locations to make predictions at.
`newB`	an optional vector of block memberships for prediction locations.
`D`	optional distance matrix for fitted and new locations specifying distances between observations.
`opts`	list of prediction options. Specified as a named list(type, num): ‘type’ specifies form of predictions, with ‘block’ (default) kriging using the block composite likelihood, ‘local’ using the closest ‘num’ points for kriging, and ‘all’ kriging with the full likelihood.
`interval`	specifing if no interval (‘none’) or a prediction interval (‘prediction’) is computed.
`level`	confidence level of prediction interval.
`nthreads`	when ‘pthreads’ are available, the number of threads to use when fitting the block composite likelihood. In most cases you will want to set this to be the number of processor cores available to you. See the ‘spacious’ manual for help enabling ‘pthreads’ support.
`gpu`	boolean to use an available ‘nVidia’ GPU for fitting full likelihood models when ‘CUDA’ support is enabled. See the ‘spacious’ manual for help enabling ‘CUDA’ support.
`engine`	defaults to engine used in fit. Can be one of ‘C’ or ‘R’. Use of ‘C’ is recommended, as ‘R’ does not have threading or GPU support is primarily for testing and prototyping.
`...`	further arguments passed to or from other methods.