Open TianningGao opened 11 months ago
Hi, Thank you for your interest in BODi. We construct a diverse random dictionary for categorical variables as well (Please see Algorithm 5 on the last page of https://arxiv.org/pdf/2303.01774.pdf). We also provided a simple illustration of the algorithm on page 18. Please check the pestcontrol
branch of this repository for our code on pestcontrol. Please let me know if that helps.
Alternatively, you could also convert the categorical variables using some encoding (one-hot / log encoding) as you mentioned.
OK, I see. I will try to implement Algorithm 5 myself and apply one-hot encoding.
Also, do you already have or plan to publish any python implementation of BODi for categorical variables? I mean the functionality to optimize, for example, the pest control problem without any binary encoding.
Hi, I think I was not fully clear but the code for categorical variables is already there in one of the branches of this repository. Please access it by this link https://github.com/aryandeshwal/BODi/tree/pestcontrol directly.
Thanks
OK, I got it. Thanks
Hello,
I have some question regarding categorical variables. Your paper for BODi shows some experiment results of the pest control probelm, but processing of non-binary categorical variables is marked as "TODO" in bodi/run_experiment.py. How did you run experiment on pest control problem with BODi? Did you use binary encoding or other encoding methods for categorical variables?