DI make.cov.obj

Creating a covariance object

We begin with help on the command line side and then present a demonstration of how to make a covariance object using the GUI display.

     Make a covariance object

DESCRIPTION:

     Create an object that stores information for the 
     covariance function.  Namely, the name of the covariance 
     function -- such as "exp.cov" -- and any other arguments 
     associated with the particular covariance function -- 
     for example, exp.cov takes theta as an argument and 
     since exp.cov uses the function rdist, it also takes 
     the logical parameter, lon.lat. 

USAGE: (Used only by a GUI dialog.  Not needed otherwise)

     make.cov.obj( temp)

REQUIRED ARGUMENTS:

temp
     temp is a list object of class "cov.obj" that 
     has all of the necessary elements for the covariance 
     object. Generally, temp is created by one 
     of: make.CMcov.obj, make.expcov.obj or make.othercov.obj.

VALUE:
     Simply returns temp without doing anything to it.

EXAMPLE:

CM.obj <- make.CMcov.obj(	mean.obj=NC.mean, 
				sd.obj=NC.sd,
				lon.lat=T,
				fit=NC.cor.fit)

EXP.obj <- make.expcov.obj(	mean.obj=NC.mean,
                                sd.obj=NC.sd,
				theta=120,
				lon.lat=T)

Other.obj <- make.othercov.obj(	name="my.cov.function",
				mean.obj=NC.mean,
                                sd.obj=NC.sd,
				cov.args=args)

The Covariance Object

To create the covariance object, click on the first toolbar button in the di toolbars or click on create covariance object under DI in the drop down menu.

Call the covariance object NC.cov.obj, by typing this in under New Covariance Object Name.

Choose a correlation model (in this case CM.cov) under Type.

Once we have chosen the type of covariance function we want to use we must click on Create Object to select the rest of the parameters.

After clicking on the Create Object button, a new dialog window should appear. This dialog window is specific to the type of covariance function selected--in our case, CM.cov. Regardless of the type of covariance function selected, we have the option of selecting mean and/or standard deviation objects. We will use NC.mean for Mean Object, NC.sd for Standard Deviation.

Since our data locations are given in longitude and latitude coordinates (using negative longitude) we must check Longitude and Latitude if it is not already checked.

Finally, since we chose CM.cov we must give it some kind of regression fit. We will use the fit NC.cor.fit, which we created in a previous example. For fit, we can use any kind of Splus function provided it accepts two vectors of the same length.

Once all the necessary parameters have been selected, click OK. You will then be back to the first dialog box. If you wish to create another covariance object, then click on apply to save the object you just created, then type a new name into the New Covariance Object Name box and create your new covariance object as above. When you are finished creating covariance objects click OK on the "Create a new covariance object" dialog.

This is software for statistical research and should not be used for commercial purposes. The authors do not guarantee the correctness of any function or program in this package.