[ANN] GModelFit.jl: a model fitting framework for Julia.
Dear community, after a few years of testing and improving the user interface, the GModelFit data-driven model fitting framework for Julia is now released!!
The package allows to define complex models (both numerical and analytical) to be compared to empirical data. A non-linear least square minimizer varies the model parameters to reduce the residuals between model and data, thus providing the best fit parameter values (with uncertainties).
Links:
- repo: GitHub - gcalderone/GModelFit.jl: A data-driven model fitting framework
- docs: Home · GModelFit.jl
Quick example:
using GModelFit
dom = Domain(0:0.01:5)
model = Model(dom, :bkg => GModelFit.OffsetSlope(1, 1, 0.1),
:l1 => GModelFit.Gaussian(1, 2, 0.2),
:l2 => GModelFit.Gaussian(1, 3, 0.4),
:main => SumReducer(:bkg, :l1, :l2))
data = GModelFit.mock(Measures, model)
best, res = fit(model, data)
The above code results in the following output:
(Components:
╭───────────┬──────────────────────────┬─────────────┬───────────┬───────────┬───────────┬─────────╮
│ Component │ Type │ Eval. count │ Min │ Max │ Mean │ NaN/Inf │
├───────────┼──────────────────────────┼─────────────┼───────────┼───────────┼───────────┼─────────┤
│ main │ SumReducer │ 204 │ 0.8851 │ 3.151 │ 1.543 │ 0 │
│ ├─ bkg │ GModelFit.OffsetSlope_1D │ 65 │ 0.8851 │ 1.398 │ 1.142 │ 0 │
│ ├─ l1 │ GModelFit.Gaussian_1D │ 87 │ 2.189e-49 │ 2.018 │ 0.2016 │ 0 │
│ └─ l2 │ GModelFit.Gaussian_1D │ 78 │ 3.608e-13 │ 1.007 │ 0.2 │ 0 │
╰───────────┴──────────────────────────┴─────────────┴───────────┴───────────┴───────────┴─────────╯
Parameters:
╭───────────┬────────┬───────┬───────────┬───────────┬────────┬───────╮
│ Component │ Param. │ Range │ Value │ Uncert. │ Actual │ Patch │
├───────────┼────────┼───────┼───────────┼───────────┼────────┼───────┤
│ bkg │ offset │ │ 0.9877 │ 0.009352 │ │ │
│ │ x0 │ │ 1 │ (FIXED) │ │ │
│ │ slope │ │ 0.1026 │ 0.004225 │ │ │
├───────────┼────────┼───────┼───────────┼───────────┼────────┼───────┤
│ l1 │ norm │ 0:Inf │ 1.01 │ 0.01883 │ │ │
│ │ center │ │ 2.002 │ 0.003727 │ │ │
│ │ sigma │ 0:Inf │ 0.1997 │ 0.003704 │ │ │
├───────────┼────────┼───────┼───────────┼───────────┼────────┼───────┤
│ l2 │ norm │ 0:Inf │ 1.002 │ 0.0288 │ │ │
│ │ center │ │ 3.005 │ 0.0102 │ │ │
│ │ sigma │ 0:Inf │ 0.3969 │ 0.01204 │ │ │
╰───────────┴────────┴───────┴───────────┴───────────┴────────┴───────╯
, Fit results:
#Data : 501 Red. fit stat.: 1.0225 (DOF: 493)
#Free : 8 Elapsed time : 0.011 s
Status: OK
)
A quick plot of data and best fit model can be obtained with (e.g.) Gnuplot.jl:
using Gnuplot
@gp data best
(this new post replaces a previous one, wrote when the package still had the old name “DataFitting”: [ANN] DataFitting.jl general purpose data fitting framework)
Comments and suggestions (as well as PRs) are very welcome!!