Salesforce Docs Portal

Overall Performance Tab for Binary Classification Use Cases

Overall Performance Tab for Binary Classification Use Cases

The Overall Performance tab shows key summary metrics for model quality.

Note
Note Einstein Discovery stories are now models. We wish we could snap our fingers to update the name everywhere, but you can expect to see the previous name in a few places until we replace it.

Navigate to the Overall Performance Tab

In Performance, click Model Evaluation, then Overall Performance.

Overall Performance tab showing accuracy analysis results

Summary Metrics

Metric Description
AUC

Area Under the Curve. Measures the logistic model’s rate of correct classification. AUC is frequently used to evaluate model quality in classification use cases.

Range:

  • 0.5 means that the model performed no better than random guessing.
  • 1.0 means that the model correctly classified the data 100% of the time. An AUC of 1.0 is suspect because it can indicate data leakage: the data used to train your model includes one or more columns that contain the information that you are trying to predict.
GINI

GINI Index. Measures how closely this classification model performs to a theoretically best possible model.

Range:

  • 0 means that the model performed no better than random guessing.
  • 1 means that the predictions matched observations exactly (and must be viewed with skepticism).
MCC

Matthews Correlation Coefficient. Measures the quality of a classification model. Provides a more even representation of the four parts of the confusion matrix than other classification metrics. In contrast, accuracy and the F1 score can be misleading when one class is predicted much more accurately than another in a classification use case.

Range:

  • -1 means that the model wrongly predicted the opposite class every time.
  • 0 means that the model performed no better than random guessing.
  • +1 means that the model correctly predicted the class every time.

Accuracy Analysis

Accuracy measures the proportion of outcomes that the model predicted correctly.

Formula:

Accuracy = (TP + TN) / Total # of Predictions

where

  • TP represents the number of true positives (positive prediction with a positive result)
  • TN represents the number of true negatives (negative prediction with a negative result)
  • Total # of Predictions that the model made (both correct and incorrect)
Metric Description
The maximum accuracy for this model is The maximum accuracy for this model.
True Positive Rate This model correctly scores records as true n% of the time.
True Negative Rate This model correctly scores records as false n% of the time.
False Positive Rate This model incorrectly scores records as true n% of the time.
False Negative Rate This model incorrectly scores records as false n% of the time.
Threshold Value The threshold value represents the tradeoff between the true positive and false positive rates. You can adjust this threshold setting in the Threshold Evaluation for Binary Classification Use Cases.

See Also