miicTeam / miic_R_package

Learning causal or non-causal graphical models using information theory
GNU General Public License v3.0
26 stars 3 forks source link

Error in discretizeMutual when matrix_u has 10 dimensions #125

Closed ylincen closed 1 year ago

ylincen commented 2 years ago

When X, Y, and Z (matrix u) are all Gaussian, and when the dimension of Z is 10, the function discretizeMutual() gives error; when the dimension of U is 5, it is fine. Screen shot can be found below.

Info: R version 4.2.1; Ubuntu 20.04.1

Thanks for the help!

Screenshot from 2022-08-29 14-41-42

Code: `rm(list = ls()) n = 1000 dz = 10

x = rnorm(n, 0, 1) y = rnorm(n, 0, 1) z = matrix(rep(0, n * dz), ncol = dz) for(i in 1:dz){ z[,i] = rnorm(n, 0, 1) }

res = discretizeMutual(X = x, Y = y, matrix_u = z, plot=F)`

franck-simon commented 2 years ago

Thank for raising this issue. From you example, we have been able to reproduce the problem and classify it as a bug. More investigation will be needed from our side to identify the causis and generate a fix