Determination of $K^0_S$ Fragmentation Functions including BESIII Measurements and using Neural Networks

In this study, we revisit the extraction of parton-to-$K^0_S$ hadron fragmentation functions, named FF24-$K^0_S$, focusing on both next-to-leading-order and next-to-next-to-leading-order accuracy in perturbative QCD. Our approach involves the analysis of single inclusive electron-positron annihilation (SIA) data. The two key improvements are, on the one hand, the incorporation of the latest experimental data from the BESIII experiment and, on the other hand, the adoption of Neural Networks in the fitting procedure. To address experimental uncertainties, the Monte Carlo method is employed. Our investigation also explores the impact of hadron mass corrections on the description of SIA data, spanning a broad kinematic regime with a particular emphasis on the range of small $z$ values. The theory prediction for $K^0_S$ production at both NLO and NNLO accuracy exhibits good agreement with experimental data within their respective uncertainties.