OpenML-CC18: Devnagari-Script
Best predictive accuracy per machine-learning flow on the Devnagari-Script classification task from the OpenML-CC18 suite.
Community results for the Devnagari-Script classification task from OpenML-CC18, OpenML's curated suite of 72 classification tasks (www.openml.org/t/167121). Each subject is a flow — a specific algorithm or pipeline — shown with the best predictive accuracy recorded for it on this task in OpenML's public evaluation listing, under the task's fixed estimation procedure.
Results
predictive_accuracy per subject over 12 measurements, ranked by average.
| Subject | Avg | Median | Min | Max | P95 | P99 | Count |
|---|---|---|---|---|---|---|---|
| keras.engine.sequential.Sequential.6CDA6D6E73373946(1) | 0.9678 | 0.9678 | 0.9678 | 0.9678 | 0.9678 | 0.9678 | 1 |
| sklearn.pipeline.Pipeline(imputation=openmlstudy14.preprocessing.ConditionalImputer,hotencoding=sklearn.preprocessing.data.OneHotEncoder,scaling=sklearn.preprocessing.data.StandardScaler,variencethre… | 0.9451 | 0.9451 | 0.9451 | 0.9451 | 0.9451 | 0.9451 | 1 |
| sklearn.pipeline.Pipeline(simpleimputer=sklearn.impute._base.SimpleImputer,standardscaler=sklearn.preprocessing.data.StandardScaler,fkceigenpro=sklearn_extra.fast_kernel.FKCEigenPro)(1) | 0.9243 | 0.9243 | 0.9243 | 0.9243 | 0.9243 | 0.9243 | 1 |
| sklearn.pipeline.Pipeline(imputation=hyperimp.utils.preprocessing.ConditionalImputer2,hotencoding=sklearn.preprocessing.data.OneHotEncoder,variencethreshold=sklearn.feature_selection.variance_thresho… | 0.8565 | 0.8565 | 0.8565 | 0.8565 | 0.8565 | 0.8565 | 1 |
| sklearn.pipeline.Pipeline(imputation=openmlstudy14.preprocessing.ConditionalImputer,hotencoding=sklearn.preprocessing.data.OneHotEncoder,variencethreshold=sklearn.feature_selection.variance_threshold… | 0.6008 | 0.6008 | 0.6008 | 0.6008 | 0.6008 | 0.6008 | 1 |
| sklearn.pipeline.Pipeline(imputer=sklearn.impute._base.SimpleImputer,estimator=sklearn.tree.tree.DecisionTreeClassifier)(1) | 0.6004 | 0.6004 | 0.6004 | 0.6004 | 0.6004 | 0.6004 | 1 |
| sklearn.pipeline.Pipeline(imputer=sklearn.impute._base.SimpleImputer,estimator=sklearn.tree.tree.DecisionTreeClassifier)(2) | 0.5993 | 0.5993 | 0.5993 | 0.5993 | 0.5993 | 0.5993 | 1 |
| sklearn.pipeline.Pipeline(imputation=sklearn.preprocessing.imputation.Imputer,estimator=sklearn.naive_bayes.GaussianNB)(1) | 0.5269 | 0.5269 | 0.5269 | 0.5269 | 0.5269 | 0.5269 | 1 |
| mlr.classif.rpart(47) | 0.1363 | 0.1363 | 0.1363 | 0.1363 | 0.1363 | 0.1363 | 1 |
| sklearn.pipeline.Pipeline(columntransformer=sklearn.compose._column_transformer.ColumnTransformer(simpleimputer=sklearn.impute._base.SimpleImputer,onehotencoder=sklearn.preprocessing._encoders.OneHot… | 0.04 | 0.04 | 0.04 | 0.04 | 0.04 | 0.04 | 1 |
| sklearn.pipeline.Pipeline(imputation=sklearn.preprocessing.imputation.Imputer,estimator=sklearn.tree.tree.DecisionTreeClassifier)(1) | 0.04 | 0.04 | 0.04 | 0.04 | 0.04 | 0.04 | 1 |
| sklearn.pipeline.Pipeline(feature_map=sklearn_extra.kernel_approximation._fastfood.Fastfood,svm=sklearn.svm.classes.LinearSVC)(1) | 0.0228 | 0.0228 | 0.0228 | 0.0228 | 0.0228 | 0.0228 | 1 |
Metrics
- predictive_accuracy — recorded on each measurement. Best predictive accuracy this flow achieved on the task: the fraction of test instances classified correctly, from 0 to 1. Higher is better
Subjects (12)
- keras.engine.sequential.Sequential.6CDA6D6E73373946(1)
- mlr.classif.rpart(47)
- sklearn.pipeline.Pipeline(columntransformer=sklearn.compose._column_transformer.ColumnTransformer(simpleimputer=sklearn.impute._base.SimpleImputer,onehotencoder=sklearn.preprocessing._encoders.OneHot…
- sklearn.pipeline.Pipeline(feature_map=sklearn_extra.kernel_approximation._fastfood.Fastfood,svm=sklearn.svm.classes.LinearSVC)(1)
- sklearn.pipeline.Pipeline(imputation=hyperimp.utils.preprocessing.ConditionalImputer2,hotencoding=sklearn.preprocessing.data.OneHotEncoder,variencethreshold=sklearn.feature_selection.variance_thresho…
- sklearn.pipeline.Pipeline(imputation=openmlstudy14.preprocessing.ConditionalImputer,hotencoding=sklearn.preprocessing.data.OneHotEncoder,variencethreshold=sklearn.feature_selection.variance_threshold…
- sklearn.pipeline.Pipeline(imputation=openmlstudy14.preprocessing.ConditionalImputer,hotencoding=sklearn.preprocessing.data.OneHotEncoder,scaling=sklearn.preprocessing.data.StandardScaler,variencethre…
- sklearn.pipeline.Pipeline(imputation=sklearn.preprocessing.imputation.Imputer,estimator=sklearn.naive_bayes.GaussianNB)(1)
- sklearn.pipeline.Pipeline(imputation=sklearn.preprocessing.imputation.Imputer,estimator=sklearn.tree.tree.DecisionTreeClassifier)(1)
- sklearn.pipeline.Pipeline(imputer=sklearn.impute._base.SimpleImputer,estimator=sklearn.tree.tree.DecisionTreeClassifier)(1)
- sklearn.pipeline.Pipeline(imputer=sklearn.impute._base.SimpleImputer,estimator=sklearn.tree.tree.DecisionTreeClassifier)(2)
- sklearn.pipeline.Pipeline(simpleimputer=sklearn.impute._base.SimpleImputer,standardscaler=sklearn.preprocessing.data.StandardScaler,fkceigenpro=sklearn_extra.fast_kernel.FKCEigenPro)(1)
Published by OpenML.
No observations in this window.