ConceptPythonBeginner
Least squares: what a fit minimizes, and why the residuals are squared
- return np.linalg.lstsq(X / sigma[:, None], y / sigma, rcond=None)[0]
- P = np.linalg.lstsq(X / sigma[:, None], (Y / sigma).T, rcond=None)[0] # all refits in one call
- coef = np.linalg.lstsq(X_poly, E, rcond=None)[0]
- p, *_ = np.linalg.lstsq(X / sigma[:, None], y / sigma, rcond=None) # rows divided by σ: weighted least squares