Struct RandomFeatures 

pub struct RandomFeatures {
    pub n_features: usize,
    pub dim: usize,
    pub weights: Vec<Vec<f64>>,
    pub biases: Vec<f64>,
}

Random Fourier features for kernel approximation.

Fields

n_features: usize

dim: usize

weights: Vec<Vec<f64>>

biases: Vec<f64>

Implementations

impl RandomFeatures

pub fn new(n_features: usize, dim: usize, seed: u64) -> Self

pub fn transform(&self, input: &[f64]) -> Vec<f64>

Transform input using random features.

pub fn kernel_approx(&self, a: &[f64], b: &[f64]) -> f64

Approximate kernel value between two inputs.