utils.types.Metrics

View Source on Github

Arize enum for metrics grouping when validating a schema column in log() call.

Use Case	SDK Metric	Metric
Regression	Metrics.REGRESSION	MAPE, MAE, RMSE, MSE, R-Squared, Mean Error
Classification	Metrics.CLASSIFICATION	Accuracy, Recall, Precision, FPR, FNR, F1, Sensitivity, Specificity
Ranking	Metrics.RANKING	NDCG
AUC & LogLoss	Metrics.AUC_LOG_LOSS	AUC, PR-AUC, Log Loss
Ranking Label	Metrics.RANKING_LABEL	GroupAUC, MAP, MRR

Method

repr()

To view applicable metrics, pass in your desired SDK Metric from above.

repr(Metrics.[SDK Metric])

Code Example

response = arize_client.log(
    model_id='sample-binary-classification-model', 
    ...
    metrics_validation=[Metrics.CLASSIFICATION]

)

response = arize_client.log(
    model_id='sample-regression-model', 
    ...
    metrics_validation=[Metrics.REGRESSION]
)

response = arize_client.log(
    model_id='sample-ranking-model', 
    ...
    metrics_validation=[Metrics.RANKING, Metrics.RANKING_LABEL]
)