Random features

AdditiveChi2Sampler([sample_steps, …])

A wrapper of sklearn.kernel_approximation.AdditiveChi2Sampler.

CompactRandomFeature([transformer_up, …])

Efficient random feature map by Compact Random Feature map.

CountSketch([n_components, degree, …])

Approximates feature map of the linear kernel by Count Sketch, a.k.a feature hashing.

FastFood([n_components, gamma, …])

Approximates feature maps of the product between random matrix and feature vectors by FastFood.

MB([n_components, dense_output])

Approximates feature map of the intersection (min) kernel by explicit feature map, which is proposed by S.Maji and A.C.Berg.

OrthogonalRandomFeature([n_components, …])

Approximates feature map of the RBF or dot kernel by Orthogonal Random Feature map.

RandomFourier([n_components, kernel, gamma, …])

Approximates feature map of the RBF kernel by Random Fourier Feature map.

RandomKernel([n_components, kernel, degree, …])

Approximates feature map of the ANOVA/all-subsets kernel by Random Kernel Feature map.

RandomMaclaurin([n_components, p, kernel, …])

Approximates feature map of a dot product kernel by Monte Carlo approximation of its Maclaurin expansion.

RandomProjection([n_components, …])

Approximates feature map of the linear product kernel by Random Projection.

SignedCirculantRandomKernel(n_components[, …])

Approximates feature map of the ANOVA kernel by Signed Circulant Random Kernel map.

SignedCirculantRandomMatrix([n_components, …])

Approximates the product between random matrix and feature vectors by signed circulant random matrix.

StructuredOrthogonalRandomFeature([…])

Approximates feature map of the RBF or dot kernel by Structured Orthogonal Random Fourier Feature map.

SubsampledRandomHadamard([n_components, …])

Approximates feature maps of the product between random matrix and feature vectors by Subsampled Randomized Hadamard Transform.

SubfeatureRandomKernel([n_components, …])

Approximates feature map of the ANOVA/all-subsets kernel by Subfeature Random Kernel Feature map.

SubfeatureRandomMaclaurin([n_components, …])

Approximates feature map of a dot product kernel by Monte Carlo approximation of its Maclaurin expansion with only sub features.

TensorSketch([n_components, degree, …])

Approximates feature map of the polynomial kernel by Tensor Sketch.

LearningKernelwithRandomFeature([…])

Learns importance weights of random features by maximizing the kernel alignment with a divergence constraint.