The outputs from singlesample.kentparams are a 3x3 matrix axes containing the estimated axes of a sample distribution, and the estimated concentration parameter kappahat and estimated ovalness parameter betahat. The inputs to distributions.kent to simulate a Kent distribution are kappa, beta, the vector mu (the mean vector of the Kent distribution), and the vector mu0 (the mean vector of the Fisher part). I can't understand why the outputs from singlesample.kentparams wouldn't match the inputs to distributions.kent, as that unnecessarily complicates the obvious use case "estimate the Kent parameters of a sample of unit vectors, then draw from a Kent distribution using the estimated parameters." I believe that the first axis from axes is the mean vector of the Kent distribution mu, but I can't figure out how to get mu0 from axes - in fact, I don't really understand what mu0 means as "the mean vector of the Fisher part" in reference to the standard form of the Kent distribution, f(x) = 1/c(kappa, beta) * exp{kappa * gamma1_dot_x + beta *(gamma2_dot_x)^2 - beta * (gamma3_dot_x)^2}. Can you please help? Thanks.
The outputs from singlesample.kentparams are a 3x3 matrix axes containing the estimated axes of a sample distribution, and the estimated concentration parameter kappahat and estimated ovalness parameter betahat. The inputs to distributions.kent to simulate a Kent distribution are kappa, beta, the vector mu (the mean vector of the Kent distribution), and the vector mu0 (the mean vector of the Fisher part). I can't understand why the outputs from singlesample.kentparams wouldn't match the inputs to distributions.kent, as that unnecessarily complicates the obvious use case "estimate the Kent parameters of a sample of unit vectors, then draw from a Kent distribution using the estimated parameters." I believe that the first axis from axes is the mean vector of the Kent distribution mu, but I can't figure out how to get mu0 from axes - in fact, I don't really understand what mu0 means as "the mean vector of the Fisher part" in reference to the standard form of the Kent distribution, f(x) = 1/c(kappa, beta) * exp{kappa * gamma1_dot_x + beta *(gamma2_dot_x)^2 - beta * (gamma3_dot_x)^2}. Can you please help? Thanks.