RIME Metrics

Module containing metric definitions for RIME Tabular.

class rime.tabular.metric.MetricName(value)

Parseable tabular metric name.

AUC = 'area_under_curve'

Binary area under the ROC curve (AUC)

MULTICLASS_AUC = 'multiclass_area_under_curve'

Multi-Class area under the ROC curve (AUC)

ACCURACY = 'accuracy'

Accuracy

MULTICLASS_ACCURACY = 'multiclass_accuracy'

Multi-class accuracy

F1 = 'f1'

Binary F1 score

MACRO_F1 = 'macro_f1'

Macro F1 score

POS_PRED_RATE = 'positive_prediction_rate'

Positive prediction rate

AVG_PRED = 'average_prediction'

Binary average prediction

AVG_CONFIDENCE = 'average_confidence'

Average confidence

ATC = 'average_thresholded_confidence'

Average Thresholded Confidence (ATC)

PRECISION = 'precision'

Binary precision

MACRO_PRECISION = 'macro_precision'

Macro precision

FPR = 'false_positive_rate'

Binary false positive rate

FNR = 'false_negative_rate'

Binary false negative rate

RECALL = 'recall'

Binary recall

MACRO_RECALL = 'macro_recall'

Macro recall

PRED_VARIANCE_POS = 'positive_prediction_variance'

Prediction variance over the positive labels

PRED_VARIANCE_NEG = 'negative_prediction_variance'

Prediction variance over the negative labels

PRED_VARIANCE_ALL = 'prediction_variance'

Prediction variance

MSE = 'mean_squared_error'

Mean squared error (MSE)

RMSE = 'root_mean_squared_error'

Root mean squared error (RMSE)

MAE = 'mean_absolute_error'

Mean absolute error (MAE)

MAPE = 'mean_absolute_percentage_error'

Mean absolute percentage error (MAPE)

RANK_CORRELATION = 'rank_correlation'

Rank correlation

MRR = 'mean_reciprocal_rank'

Mean reciprocal rank (MRR)

AVG_RANK = 'average_rank'

Average rank

AVG_ROW_WISE_PRED_DIFF = 'avg_row_wise_pred_diff'

Average row-wise prediction difference