Scaling Behavior of the Hirsch Index for Paper Citations, Failure Avalanches and Percolation Clusters

A popular measure for the success or citation inequalities of individual scientists have been the Hirsch index ($h$). If the number $n_c$ of citations are plotted against the number $n_p$ of papers (having those citations), then $h$ corresponds to the fixed point (where $n_c = h = n_p$) of the above-mentioned citation function. There have been theoretical as well as numerical studies suggesting $h\sim \sqrt{N_c} \sim \sqrt {N_p}$, for any author having $N_p = \Sigma n_p$ total papers and $N_c = \Sigma n_c$ total citations. With extensive data analysis here, we show that $h \sim \sqrt N_c /log N_c \sim \sqrt N_p /log N_p$. Our numerical study also shows that the $h$-index values for size distribution of avalanches in equal load-sharing fiber bundle models (of materials failure), consisting of $N$ fibers, also scale as $\sqrt N/ log N$. We however observe a notable discrepancy in the scaling relation of $h$ for the size distribution of clusters near the percolation point of square lattice (indicating that the scaling behavior depends on the dimension). We also show that if the number $N_s$ of the members of parliaments or national assemblies of different countries of the world are identified effectively as their Hirsch index $h$, then $N_s$ indeed scales with the population $N$ as $\sqrt N /log N$ for the respective countries and this observation may help comprehending the discrepancies with $\sqrt N$ relationship, reported in a very recent analysis.