Routines for stochastic matrices, based on mpmath

Compute the stochastic eigenvector (or stationary distribution vector) of an irreducible stochastic (or Markov) matrix, by using the Grassmann-Taksar-Heyman (GTH) algorithm, a numerically stable variant of Gaussian elimination.

• Source code
• Illustration 1: Basic usage and comparison with mpmath.eig
• Illustration 2: mp versus fp
• Illustration 3: Comparison with numpy.linalg.eig and scipy.linalg.eig
• pytest