Model Details: ============== H2OBinomialModel: gbm Model ID: GBM_1_AutoML_20200428_174837 Model Summary: number_of_trees number_of_internal_trees model_size_in_bytes min_depth 1 181 181 138744 6 max_depth mean_depth min_leaves max_leaves mean_leaves 1 6 6.00000 37 64 56.33149 H2OBinomialMetrics: gbm ** Reported on training data. ** MSE: 0.07087676 RMSE: 0.2662269 LogLoss: 0.2275785 Mean Per-Class Error: 0.1349736 AUC: 0.9553364 AUCPR: 0.888561 Gini: 0.9106729 R^2: 0.6123146 Confusion Matrix (vertical: actual; across: predicted) for F1-optimal threshold: 0 1 Error Rate 0 22902 1818 0.073544 =1818/24720 1 1540 6301 0.196404 =1540/7841 Totals 24442 8119 0.103130 =3358/32561 Maximum Metrics: Maximum metrics at their respective thresholds metric threshold value idx 1 max f1 0.400832 0.789599 200 2 max f2 0.191420 0.849698 279 3 max f0point5 0.596916 0.825186 133 4 max accuracy 0.477003 0.900341 173 5 max precision 0.997990 1.000000 0 6 max recall 0.001021 1.000000 398 7 max specificity 0.997990 1.000000 0 8 max absolute_mcc 0.447213 0.721783 183 9 max min_per_class_accuracy 0.294270 0.877694 239 10 max mean_per_class_accuracy 0.223316 0.879591 266 11 max tns 0.997990 24720.000000 0 12 max fns 0.997990 7660.000000 0 13 max fps 0.000693 24720.000000 399 14 max tps 0.001021 7841.000000 398 15 max tnr 0.997990 1.000000 0 16 max fnr 0.997990 0.976916 0 17 max fpr 0.000693 1.000000 399 18 max tpr 0.001021 1.000000 398 Gains/Lift Table: Extract with `h2o.gainsLift(, )` or `h2o.gainsLift(, valid=, xval=)` H2OBinomialMetrics: gbm ** Reported on validation data. ** MSE: 0.08841016 RMSE: 0.2973385 LogLoss: 0.2774068 Mean Per-Class Error: 0.1738575 AUC: 0.927307 AUCPR: 0.8236126 Gini: 0.854614 R^2: 0.5099851 Confusion Matrix (vertical: actual; across: predicted) for F1-optimal threshold: 0 1 Error Rate 0 11131 1304 0.104865 =1304/12435 1 934 2912 0.242850 =934/3846 Totals 12065 4216 0.137461 =2238/16281 Maximum Metrics: Maximum metrics at their respective thresholds metric threshold value idx 1 max f1 0.372001 0.722401 204 2 max f2 0.164067 0.808524 284 3 max f0point5 0.630229 0.761592 116 4 max accuracy 0.529264 0.873042 150 5 max precision 0.998128 1.000000 0 6 max recall 0.000742 1.000000 399 7 max specificity 0.998128 1.000000 0 8 max absolute_mcc 0.372001 0.632486 204 9 max min_per_class_accuracy 0.257577 0.843214 246 10 max mean_per_class_accuracy 0.207394 0.845709 266 11 max tns 0.998128 12435.000000 0 12 max fns 0.998128 3775.000000 0 13 max fps 0.000742 12435.000000 399 14 max tps 0.000742 3846.000000 399 15 max tnr 0.998128 1.000000 0 16 max fnr 0.998128 0.981539 0 17 max fpr 0.000742 1.000000 399 18 max tpr 0.000742 1.000000 399 Gains/Lift Table: Extract with `h2o.gainsLift(, )` or `h2o.gainsLift(, valid=, xval=)` H2OBinomialMetrics: gbm ** Reported on cross-validation data. ** ** 5-fold cross-validation on training data (Metrics computed for combined holdout predictions) ** MSE: 0.08874909 RMSE: 0.2979079 LogLoss: 0.2792597 Mean Per-Class Error: 0.1748988 AUC: 0.9280675 AUCPR: 0.8270713 Gini: 0.8561349 R^2: 0.5145556 Confusion Matrix (vertical: actual; across: predicted) for F1-optimal threshold: 0 1 Error Rate 0 22413 2307 0.093325 =2307/24720 1 2011 5830 0.256472 =2011/7841 Totals 24424 8137 0.132613 =4318/32561 Maximum Metrics: Maximum metrics at their respective thresholds metric threshold value idx 1 max f1 0.400627 0.729753 198 2 max f2 0.132405 0.810691 308 3 max f0point5 0.631931 0.762730 118 4 max accuracy 0.495652 0.873315 164 5 max precision 0.995777 0.998478 2 6 max recall 0.001232 1.000000 397 7 max specificity 0.997890 0.999960 0 8 max absolute_mcc 0.400627 0.642123 198 9 max min_per_class_accuracy 0.262700 0.842961 250 10 max mean_per_class_accuracy 0.233993 0.845111 261 11 max tns 0.997890 24719.000000 0 12 max fns 0.997890 7651.000000 0 13 max fps 0.000610 24720.000000 399 14 max tps 0.001232 7841.000000 397 15 max tnr 0.997890 0.999960 0 16 max fnr 0.997890 0.975768 0 17 max fpr 0.000610 1.000000 399 18 max tpr 0.001232 1.000000 397 Gains/Lift Table: Extract with `h2o.gainsLift(, )` or `h2o.gainsLift(, valid=, xval=)` Cross-Validation Metrics Summary: mean sd cv_1_valid cv_2_valid cv_3_valid cv_4_valid accuracy 0.86735666 0.0053434274 0.866421 0.8716216 0.85964376 0.8659398 auc 0.9280953 0.0044780015 0.93294436 0.9296904 0.9208378 0.9277435 aucpr 0.82710934 0.012634234 0.8447242 0.8284223 0.80986893 0.822535 err 0.13264331 0.0053434274 0.133579 0.12837838 0.14035627 0.13406019 err_count 863.8 34.8023 870.0 836.0 914.0 873.0 cv_5_valid accuracy 0.87315726 auc 0.9292603 aucpr 0.82999617 err 0.12684275 err_count 826.0 --- mean sd cv_1_valid cv_2_valid cv_3_valid cv_4_valid pr_auc 0.82710934 0.012634234 0.8447242 0.8284223 0.80986893 0.822535 precision 0.7140797 0.020288346 0.7167431 0.7198986 0.6848665 0.70805573 r2 0.51432127 0.015408624 0.5334438 0.51828897 0.49201664 0.5079919 recall 0.7498898 0.0117527945 0.7687577 0.7424837 0.75081325 0.74935895 rmse 0.2978934 0.0032907955 0.2956326 0.29425678 0.30265218 0.29938108 specificity 0.9045725 0.009438977 0.89891547 0.91128063 0.8932663 0.9026656 cv_5_valid pr_auc 0.82999617 precision 0.7408344 r2 0.51986486 recall 0.73803526 rmse 0.2975443 specificity 0.91673434