Description Usage Arguments Author(s)
A function to simulate multi-class data with a linear class-mean trend. The signal dimension is the dimension carrying all of the between-class difference, and the non-signal dimensions are noise.
1 2 3 4 5 6 7 8 9 10 11 | discr.sims.linear(
n,
d,
K,
signal.scale = 1,
signal.lshift = 1,
non.scale = 1,
rotate = FALSE,
class.equal = TRUE,
ind = FALSE
)
|
n |
the number of samples. |
d |
the number of dimensions. The first dimension will be the signal dimension; the remainders noise. |
K |
the number of classes in the dataset. |
signal.scale |
the scaling for the signal dimension. Defaults to |
signal.lshift |
the location shift for the signal dimension between the classes. Defaults to |
non.scale |
the scaling for the non-signal dimensions. Defaults to |
rotate |
whether to apply a random rotation. Defaults to |
class.equal |
whether the number of samples/class should be equal, with each
class having a prior of 1/K, or inequal, in which each class obtains a prior
of k/sum(K) for k=1:K. Defaults to |
ind |
whether to sample x and y independently. Defaults to |
Eric Bridgeford
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.