Inverse Ising techniques to infer underlying mechanisms from data
Zeng Hong-Li1, 2, †, Aurell Erik3, 4, ‡
       

Mean square error (ε) versus (a) data length L, (b) system size N, (c) external field θ and (d) temperature 1/g. Black squares show nMF, red circles, TAP, blue up triangle SHO and pink down triangle AVE respectively. The parameters are g = 0.3, N = 20, θ = 0, L = 107 except when varied in a panel.