以下分箱规则应添加到 Simone 的列表中,这些规则已变得更加普遍:
假设互信息是由它们的联合熵调整的边际熵的总和,I(X,Y)=H(X)+H(Y)−H(X,Y)
以及的最佳分箱规则 是H(X)H(Y)Hacine-Gharbi et al. (2012)
BX=round(ξ6+23ξ+13)
其中ξ=(8+324N+1236N+729N2−−−−−−−−−−−√)13
而的最优分箱规则 为H(X,Y)Hacine-Gharbi and Ravier (2018)
BX=BY=round[12–√(1+1+24N1−ρ2−−−−−−−−−√)12]
在测量的各个项时应用这些分箱规则,您应该有一个互信息的最佳分箱低偏差估计器。I(X,Y)=H(X)+H(Y)−H(X,Y)
Hacine-Gharbi, A., and P. Ravier (2018): “A Binning Formula of Bi-histogram
for Joint Entropy Estimation Using Mean Square Error Minimization.”
Pattern Recognition Letters, Vol. 101, pp. 21–28.
Hacine-Gharbi, A., P. Ravier, R. Harba, and T. Mohamadi (2012): “Low Bias
Histogram-Based Estimation of Mutual Information for Feature Selection.”
Pattern Recognition Letters, Vol. 33, pp. 1302–8.