search

AI4LIFE-GROUP / OpenXAI

OpenXAI : Towards a Transparent Evaluation of Model Explanations
https://open-xai.github.io/
MIT License
232 stars 39 forks source link

Fix calculation for log standard error #31

Open statkwon opened 2 months ago

statkwon commented 2 months ago

I created this PR to fix the code for calculating log standard error.

Before

log_mu, log_std = np.log(mean_score), np.log(std_err)

After

log_mu = np.log(mean_score)
log_std = np.std(np.log(score)) / np.sqrt(len(score))