Collate, fit, and plot dimuon cross-sections from Whizard 3.0.3
Read in data files and fit i) ALR vs ECM and ii) unpolarized cross-section (in units of R) vs ECM.
Result is plots of ALR, xsLR, xsRL, xsU, Rmumu vs ECM with the superimposed parametrizations. (xsLR, xsRL, xsU models are inferred from the ALR and Rmumu fits using the QED point-like cross-section).
Current fits are a polynomial with 3 free parameters to ALR, and a fit to Rmumu using a BW and polynomial terms with 4 free parameters. See mymodels.py, fitconfig,py and fitting.py for details.
The uncertainties in the Whizard cross-section have been scaled up a little using the chi-squared amongst the various iterations. I have seen some occurrences of the Whizard cross-section integration being quite discrepant, so I think there is still an expectation of some outliers and not particularly well modeled uncertainties. The run tests, while nominally less powerful (than chi-squared) look encouraging.
ALR fit. Chi-squared / dof = 38.0/24, pvalue = 3.4% Run-test p-value = 88%
Rmumu fit. Chi-squared / dof = 65.2/23, pvalue = small Run-test p-value = 37%
The 6 figures with superimposed models. Note only Figure 4 and Figure 6 are directly fitted.
Logfile from running the code