Create a network object.

DESCRIPTION:

     A list object that contains the locations, grid points 
     (for prediction) and covariance object used with Di.

USAGE:
     make.network.obj(locs, grid=NULL, cov=cov.obj, [Other Options])

REQUIRED ARGUMENTS:

locs
     an n X 2 matrix of locations.
     Usually lon and lat locations.

cov
     the name of a covariance object.
     Usually one created by make.cov.obj.

OPTIONAL ARGUMENTS:

grid
     Grid points used for prediction.

nx, ny
     The number of x and/or y grid points respectively.  
     Default is 60 for each.

xlower, xupper, ylower and yupper
     Ranges for the grid.  The default is to use the 
     ranges of the locations object (locs).

VALUE:
     List object.

SEE ALSO:
     make.cov.obj, network.plotter, edit.network

EXAMPLES:

     NC.net <- make.network.obj(	locs=NC.ex$loc,
					Yname="NC.ozmax8",
					cov="NC.cov.obj",
					nx=50,
					ny=80)

To create the network object click on the second toolbar button of the DI toolbar buttons or Make Network Object from the DI drop down menu.

Return Value
We will call our new network object NC.net by typing this in the Network Object box under Return Value.

Locations
We type NC.ex$loc in the locations list box. This will not appear in the Locations drop down menu because it is an n X 2 matrix, but it is hidden within a list object of its own. The locations object you choose must be an n X 2 matrix.

Measurements
Even though our list object NC.ex has the measurements component, NC.ex$ozmax8, we must use the object, NC.ozmax8, which we created from NC.ex$ozmax8 in the Measurements box. The way in which the GUI handles the measurements matrix prohibits it from understanding list objects. Note that our ozmax8 dataset will not likely be part of the drop down menu either; this has to do with it having originally been a data frame.

Generally, the Measurements matrix can be any matrix object, but the number of columns must match the number of rows of the locations object; each row represents a single set of readings from the network. It is a good idea to make sure the row names of the Measurements matrix correspond to what each instance (row) represents; in this case as dates. If no row names are given, then Di will use row numbers as a default for the row names.

Covariance Object
Here we select the covariance object NC.cov.obj; created in the make.cov.obj example, and hence should be listed as an option in the Covariance Object drop down menu. Note that we must create the covariance object before we create the network object.

Grid
For the grid, we could use a same grid we already created by typing the name of the grid we created (which may even be part of the drop down menu if it has been assigned the class "grid") into the Grid Points for Prediction box. Alternatively, here we chose to type 50 in for number of x grid points and 80 in for number of y grid points. If we do not choose anything, Di will create a 60 X 60 grid based on the x and y ranges of the locations.

Click OK and the new network object will be created.

[an error occurred while processing this directive]