Initialize worse_model_count if model accuracy is better in BoostAODE
This commit is contained in:
@@ -14,6 +14,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
|
||||
### Changed
|
||||
|
||||
- Sample app now is a separate target in the Makefile and shows how to use the library with a sample dataset
|
||||
- The worse model count in BoostAODE is reset to 0 every time a new model produces better accuracy, so the tolerance of the model is meant to be the number of **consecutive** models that produce worse accuracy.
|
||||
|
||||
## [1.0.4] 2024-03-06
|
||||
|
||||
|
Reference in New Issue
Block a user