Determination of non-perturbative corrections in diphoton production at the LHC using iterative methods

This thesis investigates non-perturbative corrections for diphoton events from pp collisions at the LHC from parton level to hadron level. Iterative unfolding based on Bayes' theorem, as proposed by D'Agostini, is used on fake diphoton parton level samples. The results are compared to bin-by-bin unfolding. Influences of the diphoton identification cut parameters on the results are tested. Diphoton events are simulated using Sherpa 2.2.0 and analyzed using Rivet 2.4.1. Unfolding was performed using the ROOT framework RooUnfold. Non-perturbative corrections are primarily observed as efficiencies. Smearing effects are low for the investigated observables. The unfolding results show improved accuracy of iterative unfolding for one or two iterations over bin-by-bin corrections. This is not influenced when varying the cut parameters in the diphoton identification analysis. However, limitations of comparing the unfolded results to the hadron level distributions of independent test samples show, that further investigation is necessary.