ArcCosine

class gpjax.kernels.ArcCosine(active_dims=None, order=0, variance=1.0, weight_variance=1.0, bias_variance=1.0, n_dims=None, compute_engine=<gpjax.kernels.computations.dense.DenseKernelComputation object>)

Bases: AbstractKernel

The ArCosine kernel.

This kernel is non-stationary and resembles the behavior of neural networks. See Section 3.1 of [Cho and Saul (2011)](https://arxiv.org/abs/1112.3712) for additional details.

Parameters:

• active_dims (list[int] | slice)

• order (Literal[0, 1, 2])

• variance (AbstractUnwrappable)

• weight_variance (AbstractUnwrappable)

• bias_variance (AbstractUnwrappable)

• n_dims (int | None)

• compute_engine (AbstractKernelComputation)