mxfusion.components.distributions.gp.kernels.rbf

Members

class mxfusion.components.distributions.gp.kernels.rbf.RBF(input_dim, ARD=False, variance=1.0, lengthscale=1.0, name='rbf', active_dims=None, dtype=None, ctx=None)

Bases: mxfusion.components.distributions.gp.kernels.stationary.StationaryKernel

Radial Basis Function kernel, aka squared-exponential, exponentiated quadratic or Gaussian kernel:

\[k(r^2) = \sigma^2 \exp \bigg(- \frac{1}{2} r^2 \bigg)\]
Parameters:
  • input_dim (int) – the number of dimensions of the kernel. (The total number of active dimensions)
  • ARD (boolean) – a binary switch for Automatic Relevance Determination (ARD). If true, the squared distance is divided by a lengthscale for individual dimensions.
  • variance (float or MXNet NDArray) – the initial value for the variance parameter (scalar), which scales the whole covariance matrix.
  • lengthscale (float or MXNet NDArray) – the initial value for the lengthscale parameter.
  • name (str) – the name of the kernel. The name is used to access kernel parameters.
  • active_dims ([int] or None) – The dimensions of the inputs that are taken for the covariance matrix computation. (default: None, taking all the dimensions).
  • dtype (numpy.float32 or numpy.float64) – the data type for float point numbers.
  • ctx (None or mxnet.cpu or mxnet.gpu) – the mxnet context (default: None/current context).