Fig. 4: Histograms of frequency estimators.
From: Power-law scaling of correlations in statistically polarised nano-NMR
The data is obtained from non-linear least squares fitting of auto-correlations for experimental (upper row) and numerical (lower row) Qdyne time-traces. In blue, fitting function assumes exponential decay of correlations while in orange the power-law model is considered. Each estimator corresponds to the highest R2 in 500 fittings of the same time-trace. In (a), the phase φ is left as a free fitting parameter while in (b) it is kept fixed to φ = 0. (c) displays the results for numerical data generated for auto-correlations with power-law decay while in (d) the data is generated for exponentially decaying auto-correlations.
