02.26.2021
New Releases, Enhancements, Changes + Arize in the News!
For models that have a score/class combination, we now allow sending in both values as the prediction/ Teams that were previously using two models (one with prediction score, one with prediction class) or passing in the prediction score as a feature can now send one model with both score and class.
All classification metrics are supported for this model type in your Arize dashboards:
- Easily analyze models that have both a prediction score and prediction label. Metrics can be built for both prediction values.
- Scored models allow users to track drift on both the prediction label and the prediction score.
- Improves usability by combining the probability score and prediction class. This can be used in segmenting prediction and cohort analysis.
Last modified 2yr ago