Open ghisloine opened 9 months ago
Hi - To compute the complementarity metric, you can use Eq. 12 of the paper. (The parameters \rho
and \alpha
can be found in the output files after running Shopper.)
Okay, but there are two files for each calculation which are mean and std. I am not good at probability. So when I try to run with the example dataset there are 2 folders for each rho and alpha. Should I calculate Eq. 12 using param_xxx_mean files using columns value1, value2 ...? Or how can I use std and mean files together?
Hi - The simplest approach is to ignore all "std" files and use "param_xxx_mean" only (using all columns, as you describe).
I am a bit confused about how can i calculate complementary items of given item ?