covarmat: gpe covariance matrix class

In goldingn/gpe: Gaussian Process Everything

Description

Generic functions associated with the covarmat class.

Usage

is.covarmat(x)

## S3 method for class 'covarmat'
image(x, axes = TRUE, legend = TRUE,
  col = palPuRd(100), ...)

palPuRd(n)

Arguments

x	an object of class covarmat, or an object to be tested as one.
axes	whether to add axes to the image plot of x.
legend	whether to add a legend to the image plot of x.
col	colour table used to plot the image
n	length of the continuous colour vector required
...	other arguments to be passed to lower-level functions, such as image.

Details

is.covarmat returns a logical indicating whether the object is a gpe covarmat object and image.covarmat provides a nice plot of the values in the covariance matrix; in the style of, and depending on image.plot. palPuRd uses the PuRd color palette from RColorBrewer to produce a non-divergent colour palette for visualising these matrices.

Examples

 
# construct a kernel with one feature
k1 <- rbf('a')

# create a fake dataframe
df <- data.frame(a = sort(runif(50, -3, 3)))
 
# evaluate the kernel
K <- k1(df)


# is it a covarmat object? 
is.covarmat(K)
 

# visualise it 
image(K)

# create a new dataframe
df2 <- data.frame(a = sort(runif(100, -3, 3)))

# visualise the covariance matrix between the two dataframes
image(k1(df, df2))


# visualise the covariance matrix between these dataframes using
# the purple-red colour scheme (actually the default) but with fewer levels
image(k1(df, df2), col = palPuRd(10))

goldingn/gpe documentation built on May 17, 2019, 7:41 a.m.