Per-detector F1 on synthetic warehouse shapes, NAB subset, and Yahoo Webscope S5 subset. results.csv ยท algorithm docs
| Detector | Family | F1 v | Precision | Recall | N | |
|---|---|---|---|---|---|---|
| bocpd | timeseries | 1.000 | 1.000 | 1.000 | 1 | 100% |
| callable_check | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| cardinality_in_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| column_pair_comparison | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| completeness | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| composite_uniqueness | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| cusum | timeseries | 1.000 | 1.000 | 1.000 | 1 | 100% |
| date_format | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| date_part_missing_fraction | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| double_mad_outlier_fraction | outlier | 1.000 | 1.000 | 1.000 | 1 | 100% |
| ecod | outlier | 1.000 | 1.000 | 1.000 | 1 | 100% |
| freshness_seconds_behind | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| generalized_esd | outlier | 1.000 | 1.000 | 1.000 | 1 | 100% |
| hbos | outlier | 1.000 | 1.000 | 1.000 | 1 | 100% |
| iqr_fence | outlier | 1.000 | 1.000 | 1.000 | 1 | 100% |
| js_divergence | drift | 1.000 | 1.000 | 1.000 | 1 | 100% |
| kl_divergence | drift | 1.000 | 1.000 | 1.000 | 1 | 100% |
| ks_pvalue | drift | 1.000 | 1.000 | 1.000 | 1 | 100% |
| lof | outlier | 1.000 | 1.000 | 1.000 | 1 | 100% |
| mahalanobis_distance | outlier | 1.000 | 1.000 | 1.000 | 1 | 100% |
| max_in_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| median_in_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| min_in_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| mmd | drift | 1.000 | 1.000 | 1.000 | 1 | 100% |
| monotonicity | timeseries | 1.000 | 1.000 | 1.000 | 1 | 100% |
| null_fraction | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| numeric_mean | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| outlier_fraction_drift | drift | 1.000 | 1.000 | 1.000 | 1 | 100% |
| page_hinkley | timeseries | 1.000 | 1.000 | 1.000 | 1 | 100% |
| quantile_in_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| referential_integrity_rate | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| regex_match | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| row_count_in_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| schema_change | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| set_exclusion | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| set_membership | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| sql_assertion_violation | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| stddev_in_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| string_case_violation | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| string_length_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| sum_in_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| uniqueness | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| validity | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| value_in_range | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| volume | rule | 1.000 | 1.000 | 1.000 | 1 | 100% |
| wasserstein_1 | drift | 1.000 | 1.000 | 1.000 | 1 | 100% |
| auto_outlier | outlier | 0.952 | 0.909 | 1.000 | 1 | 95% |
| benford_law_fit | distribution | 0.952 | 0.909 | 1.000 | 1 | 95% |
| grubbs | outlier | 0.952 | 0.909 | 1.000 | 1 | 95% |
| zscore_outlier_fraction | outlier | 0.952 | 0.909 | 1.000 | 1 | 95% |
| adjusted_boxplot_fraction | outlier | 0.909 | 0.833 | 1.000 | 1 | 91% |
| adwin | drift | 0.909 | 0.833 | 1.000 | 1 | 91% |
| psi | drift | 0.769 | 0.625 | 1.000 | 1 | 77% |
| isolation_forest_fraction | outlier | 0.741 | 0.588 | 1.000 | 1 | 74% |
| holt_winters | timeseries | 0.667 | 0.500 | 1.000 | 1 | 67% |
| mutual_information | distribution | 0.667 | 0.500 | 1.000 | 1 | 67% |
| one_class_svm | outlier | 0.667 | 0.500 | 1.000 | 1 | 67% |
| stl_residual_zscore | timeseries | 0.667 | 0.500 | 1.000 | 1 | 67% |
| matrix_profile | timeseries | 0.182 | 1.000 | 0.100 | 1 | 18% |
| chi_square_drift | drift | 0.000 | 0.000 | 0.000 | 1 | 0% |
| cramers_v | distribution | 0.000 | 0.000 | 0.000 | 1 | 0% |
| mad_outlier_fraction | outlier | 0.000 | 0.000 | 0.000 | 1 | 0% |
| prophet_anomaly | timeseries | 0.000 | 0.000 | 0.000 | 1 | 0% |
| remote_check | rule | 0.000 | 0.000 | 0.000 | 1 | 0% |