Fit a Physics-Informed Quantum Kernel Ridge Regression

Composes piml_quantum_kernel() with the closed-form KRR dual solution \(\boldsymbol{\alpha} = (K + \lambda I)^{-1} y\). The returned object carries the training-time residuals and ODE fit so predict() handles the full forward pipeline (ODE predict -> residual -> PI kernel row -> dual sum).

Usage

piml_qkrr_fit(
  X,
  y,
  depths,
  ode_fit,
  alpha = 0.7,
  sigma = NULL,
  reps = 2L,
  lambda = 0.1,
  backend = c("rcpp", "r")
)

Arguments

X, y, depths, ode_fit, alpha, sigma, reps, backend

See piml_quantum_kernel().

lambda

Ridge regulariser; positive.

Value

An edaphos_piml_qkrr fit.