search

tensorflow / addons

Useful extra functionality for TensorFlow 2.x maintained by SIG-addons
Apache License 2.0
1.69k stars 613 forks source link

Add new metrics - Matthews correlation coefficient #457

Closed autoih closed 5 years ago

autoih commented 5 years ago

Most datasets are unbalanced (different sizes of positive and negative samples); Although F1 score is usually considered as metric; however, it needs to be careful to assign positive samples. If assign in a wrong way, the score will be totally different. One advantage of MMC (Matthews correlation coefficient) can avoid this kind of situation. I'd like to know if it will be considered as an important metric. Thanks.

Squadrick commented 5 years ago

@autoih Do you want to contribute to Addons by opening a PR?

Also, in the future, we would prefer if you used the template for feature requests.

autoih commented 5 years ago

Thanks @Squadrick. I'd love to contribute this. Yes, I'll use the template in the future.

Squadrick commented 5 years ago

@autoih Looking forward to the PR.

Squadrick commented 5 years ago

Reference: Wikipedia