The first iteration of our ML model is done and working. However, while it works well for XS-leads, the accuracy for classes is not quite good yet due to the significant imbalance. This item aims to fix that circumstance by improving the prediction accuracy not only but especially for those classes.
Notes
Improving the (computational) efficiency of the model or other aspects is also fine as part of this item.
For even finer model / parameter tuning, a continuation story will most likely be created next sprint.
Optional: Consider working with an experiment canvas as a form of documentation.
Acceptance Criteria
Several attempts have been tried and documented to improve the model
Attempts have been made and documented to make scores across different classes consistent
The first iteration of our ML model is done and working. However, while it works well for XS-leads, the accuracy for classes is not quite good yet due to the significant imbalance. This item aims to fix that circumstance by improving the prediction accuracy not only but especially for those classes.
Notes
Acceptance Criteria