Closed scheidan closed 1 year ago
After some discussion Carlo and I came to the following, preliminary conclusions.
We get multiple distances (rho) from f_dist. We change the initialization as follows:
f_dist
n_particle
distances
u
epsilon_init
exp(- u ./ epsilon_init)
We need to think about:
The multivariate CDF is not a correct solution!
After some discussion Carlo and I came to the following, preliminary conclusions.
We get multiple distances (rho) from
f_dist
. We change the initialization as follows:n_particle
prior samples and compute distancesdistances
-> this function maps a vector of summary distances to the scalar numberu
.epsilon_init
(asepsilon_init
relates now tou
it shoud have an easy interpretation)exp(- u ./ epsilon_init)
until we haven_particle
samples.We need to think about: