This article describes the model overview screen and how to use the sparkline graph as a guideline to the ongoing health of a deployed model.
In this article:
The dashboard lists each deployed model as a tile. A tile contains a snapshot of the prediction results for that model. Click a tile to view the details of a that model. To access the dashboard view, go to Predict > Overview.
A tile may display the following information about a model:
Strength ratings are recalculated automatically and the sparkline graph reflects live performance metrics.
In the dashboard view, the strength score displays in parentheses as a decimal number, such as (0.91). This number represents the F1 score that the rating is based on.
The strength rating cannot be known or calculated for a deployed model until the initial prediction timeframe has elapsed. For this reason, the strength rating of a model is not available until the day after the prediction timeframe. For example, if a model to predict the likelihood to purchase in the next 10 days is deployed on January 1st, the rating cannot be calculated until after January 10th.
Click a tile to display an expanded view of the model. In the expanded view, the sparkline graph shows the trend of the prediction ranges over time.
The prediction ranges are based on the following strength scores:
The sparkline graph adapts to versioning. For example, if you first deploy "Version 1" and later undeploy "Version 1" and deploy "Version 2", the graph automatically switches from "Version 1" to "Version 2". The adaptive chart ensures that you see the strength of the output attribute for the latest deployed version of the model.
See the following articles for additional information about model scores and evaluation: