Non-parametric profile gradient estimation

Reliable profile and profile gradient estimates are of utmost importance for many different physical models in fusion science, e.g. transport modeling or mode stabilization. Fitting profiles to a collection of results from different diagnostics defines basic work in plasma physics. The fitting results often crucially depend on the functional representation of the profile. In particular, the estimation uncertainty of the profile and, even worse, the estimation of the profile gradient and its uncertainty is closely coupled withthe providedprofileflexibility.This is onereason why profile gradient uncertainties are usually not estimated. Profile flexibility to allow for a formfree description of the data often competes with profile reliability. The estimation reliability decreases with increasing number of degree of freedom provided. The problem of the proper choice of the functional representation of the profile is hampered by measurement errors of the usually pointwise measurements of profiles and by lack of information in profile segments. Severe complications arise from systematic deviations due to inconsistent diagnostics. The aim is to have a robust technique to allow for a reasonable balance between flexibility and reliability. Flexibility is frequently obtained by using non-parametric profile functionals, e.g. linear interpolation between pointwise estimations or cubic or B-splines. The reliability of