Imbalanced targets - Githubissues

angelmtenor / data-science-keras

Data science projects with Keras

MIT License

3 stars 4 forks source link

Imbalanced targets #3

Closed angelmtenor closed 1 year ago

angelmtenor commented 7 years ago

Use f1 score (or generalized f score)
Use stratified train test split
use sample_weight / class_weight (sklearn, keras)
Modify cutoff value?, Over-sample minority classes?

angelmtenor commented 1 year ago

Solved in new repos