Module lm 

Levenberg-Marquardt least-squares optimizer.

Minimizes χ² = Σᵢ [(y_obs - y_model)² / σᵢ²] by iteratively solving:

(JᵀWJ + λ·diag(JᵀWJ)) · δ = JᵀW·r

where J is the Jacobian, W = diag(1/σ²), r = y_obs - y_model, and λ is the damping parameter.

SAMMY Reference

• fit/ module, manual Sec 4 (Bayes equations / generalized least-squares)

Structs

FlatMatrix

Row-major flat matrix for cache-friendly storage.

LmConfig

Configuration for the LM optimizer.

LmResult

Result of a Levenberg-Marquardt fit.

Traits

FitModel

A model function that can be fitted.

Functions

levenberg_marquardt

Run the Levenberg-Marquardt optimizer.