dotnet / dotnet/machinelearning-modelbuilder
Adjustable regularization
Open
Feature request
Stale
- Dominant language
- Dockerfile
- Stars
- 285
- Forks
- 66
- PR merge metrics
- No merged PRs in 30d
Description
**Is your feature request related to a problem? Please describe.**
I know that my training set for a classification problem has only 80% accurate labels. Today the models produced by modelbuilder are trying to overfit. I'd like to train a regularized model that doesn't overfit.
**Describe the solution you'd like**
I would like the ability to control the regularization strength/type used to train the model.
**Describe alternatives you've considered**
None
**Additional context**
-
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.