Hi, I am trying to contstruct a mutual informtation matrix using the function build.mim() from the minet package: > build.mim(myData, estimator="empirical") and I have two questions. 1.) The estimator "empirical" computes the entropy of the empirical probability distribution. This probably means it is using this definition of the entropy: H = sum_x sum_y p(x,y)* log( p(x,y)/ (p(x)*p(x)) ). Is this correct? 2.) Which is basis is used for the log? Thanks a lot for your in advance. Bettina Keller