Test of Levenberg-Marquardt nonlinear least squares algorithms
- Results from known Jacobian (the first derivative) of the function
Samples with parameters (5.0, 0.1, 1.0)
With Jacobian
Levenberg-Marquardt returned in 14 iter, reason 2, sumsq 0.441433 [137.625]
Best fit parameters: 4.957659 0.09761961 0.986824
Without Jacobian
Levenberg-Marquardt returned in 59 iter, reason 2, sumsq 0.441433 [137.625]
Best fit parameters: 4.957659 0.0976196 0.986824
Samples with parameters (5.0, 0.1, 1.0)
With Jacobian
Levenberg-Marquardt returned in 15 iter, reason 2, sumsq 0.447505 [136.774]
Best fit parameters: 4.98591 0.09484077 0.9286256
Without Jacobian
Levenberg-Marquardt returned in 95 iter, reason 2, sumsq 0.447505 [136.774]
Best fit parameters: 4.98591 0.09484074 0.9286256
sudo apt-get install libblas-dev liblapack-dev