search

seorim0 / ResUNet-LC

2D residual U-Net (ResUNet) and a lead combiner (LC) for 12-lead ECG Abnormality Classification
MIT License
8 stars 0 forks source link

What is the threshold of F1-score, I am super appreciative. #2

Closed poult-lab closed 1 month ago

poult-lab commented 1 month ago

Hello author,

I am wondering what threshold you set for the F1-score. I saw in the codebase that you used sklearn.metrics.average_precision_score() to get the AUPRC. We need to set a threshold to calculate the F1-score using sklearn.metrics.f1_score(). When the output is larger than the threshold, it is set to 1; otherwise, it is set to 0. Could you share this threshold with us?

seorim0 commented 1 month ago

Hello! Thank you for your interest in our project.

Two methods were used to determine the threshold. 1) Using the validation dataset, increase the threshold from 0 to 1 in increments of 0.01 and specify the value that produces the best F1-score as the threshold. 2) Fixed at 0.5

Experimentally, there was no significant performance difference between the two methods.

poult-lab commented 1 month ago

Thank you, author, for your prompt reply.