Event Timeline
Comment Actions
Damaging:
(p3)ladsgroup@ores-compute-01:~/editquality$ make models/plwiki.damaging.gradient_boosting.model cat datasets/plwiki.features_damaging.20k_2016.tsv | cut -f2- | \ revscoring train_test \ revscoring.scorer_models.GradientBoosting \ editquality.feature_lists.plwiki.damaging \ --version=0.1.1 \ -p 'max_depth=5' \ -p 'learning_rate=0.01' \ -p 'max_features="log2"' \ -p 'n_estimators=700' \ -s 'table' -s 'accuracy' -s 'precision' -s 'recall' -s 'pr' -s 'roc' -s 'recall_at_fpr(max_fpr=0.10)' -s 'filter_rate_at_recall(min_recall=0.90)' -s 'filter_rate_at_recall(min_recall=0.75)' \ --balance-sample-weight \ --center --scale \ --label-type=bool > \ models/plwiki.damaging.gradient_boosting.model 2016-07-03 21:29:52,824 INFO:revscoring.utilities.train_test -- Training model... 2016-07-03 21:30:13,456 INFO:revscoring.utilities.train_test -- Testing model... ScikitLearnClassifier - type: GradientBoosting - params: max_leaf_nodes=null, min_samples_split=2, scale=true, balanced_sample=false, subsample=1.0, presort="auto", min_samples_leaf=1, verbose=0, min_weight_fraction_leaf=0.0, learning_rate=0.01, max_depth=5, balanced_sample_weight=true, init=null, max_features="log2", center=true, n_estimators=700, random_state=null, warm_start=false, loss="deviance" - version: 0.1.1 - trained: 2016-07-03T21:30:13.453169 Table: ~False ~True ----- -------- ------- False 4064 222 True 20 52 Accuracy: 0.944 Precision: 0.19 Recall: 0.722 PR-AUC: 0.291 ROC-AUC: 0.931 Recall @ 0.1 false-positive rate: threshold=0.972, recall=0.014, fpr=0.0 Filter rate @ 0.9 recall: threshold=0.11, filter_rate=0.783, recall=0.903 Filter rate @ 0.75 recall: threshold=0.382, filter_rate=0.924, recall=0.75
Comment Actions
Good faith:
(p3)ladsgroup@ores-compute-01:~/editquality$ make models/plwiki.goodfaith.gradient_boosting.model cat datasets/plwiki.features_goodfaith.20k_2016.tsv | cut -f2- | \ revscoring train_test \ revscoring.scorer_models.GradientBoosting \ editquality.feature_lists.plwiki.goodfaith \ --version=0.1.1 \ -p 'max_depth=3' \ -p 'learning_rate=0.01' \ -p 'max_features="log2"' \ -p 'n_estimators=700' \ -s 'table' -s 'accuracy' -s 'precision' -s 'recall' -s 'pr' -s 'roc' -s 'recall_at_fpr(max_fpr=0.10)' -s 'filter_rate_at_recall(min_recall=0.90)' -s 'filter_rate_at_recall(min_recall=0.75)' \ --balance-sample-weight \ --center --scale \ --label-type=bool > \ models/plwiki.goodfaith.gradient_boosting.model 2016-07-04 02:46:35,358 INFO:revscoring.utilities.train_test -- Training model... 2016-07-04 02:46:47,488 INFO:revscoring.utilities.train_test -- Testing model... ScikitLearnClassifier - type: GradientBoosting - params: min_weight_fraction_leaf=0.0, max_depth=3, warm_start=false, n_estimators=700, init=null, random_state=null, max_features="log2", max_leaf_nodes=null, min_samples_split=2, min_samples_leaf=1, learning_rate=0.01, balanced_sample=false, balanced_sample_weight=true, scale=true, subsample=1.0, presort="auto", loss="deviance", center=true, verbose=0 - version: 0.1.1 - trained: 2016-07-04T02:46:47.484697 Table: ~False ~True ----- -------- ------- False 26 3 True 213 4116 Accuracy: 0.95 Precision: 0.999 Recall: 0.951 PR-AUC: 1.0 ROC-AUC: 0.979 Recall @ 0.1 false-positive rate: threshold=0.034, recall=1.0, fpr=0.007 Filter rate @ 0.9 recall: threshold=0.741, filter_rate=0.105, recall=0.9 Filter rate @ 0.75 recall: threshold=0.886, filter_rate=0.255, recall=0.75