pub fn linear_least_squares(
a: &[Vec<f64>],
b: &[f64],
bounds: Bounds,
tol: f64,
limits: Limits,
statistical_assumptions: bool,
) -> Result<LeastSquaresResult, Error>Expand description
Active-set bounded linear least squares over the shared, pivot-free Jacobi SVD.
Each iteration solves the reduced problem for the currently free variables, fixes the first bound encountered, and releases a bound only when its KKT multiplier has the wrong sign. This is a bounded-variable least-squares path, not an unconstrained solve followed by clipping.