search

houchengbin / OpenANE

OpenANE: the first Open source framework specialized in Attributed Network Embedding. The related paper was accepted by Neurocomputing. https://doi.org/10.1016/j.neucom.2020.05.080
MIT License
42 stars 6 forks source link

parameter searching #7

Open neenerrh opened 3 years ago

neenerrh commented 3 years ago

Hello. is there any tool for searching for the optimal parameters or does it have to be done manually

houchengbin commented 3 years ago

Hi. You may write a .sh file to conduct the grid search. An example of that could be https://github.com/houchengbin/GloDyNE/blob/master/bash/NE_AS733.sh And to run the .sh file, please follow https://github.com/houchengbin/GloDyNE#advanced-for-batch-testing

What you may need to modify are those parameters specified in https://github.com/houchengbin/OpenANE/blob/144b882dffc6aa68c0a482383883a6a97013371b/src/main.py#L25 For example, to search "ABRW-beta" for the method presented in our paper, you could write the following commands in a .sh file.

python src/main.py --method abrw --task lp_and_nc --emb-file emb/cora_abrw_emb --save-emb --ABRW-beta 0.1
python src/main.py --method abrw --task lp_and_nc --emb-file emb/cora_abrw_emb --save-emb --ABRW-beta 0.2
.
.
.
python src/main.py --method abrw --task lp_and_nc --emb-file emb/cora_abrw_emb --save-emb --ABRW-beta 0.9

Let me know if you have further problems.

neenerrh commented 3 years ago

Thank you very much. I will give it a try