- 🛠 [PUBDEV-8226] - Fixed h2odriver invalid argument error on Java 11.
- [PUBDEV-8224] - Fixed GLM
- ⬆️ [PUBDEV-8219] - Upgraded to latest version of Javassist (3.28).
- 🛠 [PUBDEV-8207] - Fixed H statistic gpu assertion error.
- 🛠 [PUBDEV-8195] - Fixed predict contributions failure in multi-MOJO environments.
- 🛠 [PUBDEV-8194] - Fixed bug in ordinal GLM class predictions.
- 🛠 [PUBDEV-8188] - Fixed Partial Dependent Plot not working with Flow.
- ⚡️ [PUBDEV-8181] - Updated to current Python syntax.
- 🛠 [PUBDEV-8167] - Fixed bug in ordinal GLM class predictions.
- 👍 [PUBDEV-8141] - Added support for refreshing HDFS delegation tokens for standalone H2O.
🆕 New Feature
- [PUBDEV-8109] - Obtained Friedman’s H statistic for XGBoost and GBM.
- ⚠ [PUBDEV-8150] - Added a warning message when using
alphaas a hyperparameter for GLM
- [PUBDEV-8136] - Fixed the printing for
pr_aucin cross-validation summaries.
🆕 New Feature
- 🐎 [PUBDEV-8131] - Added parameter
auc_typeto performance method to compute multiclass AUC.
- ⬆️ [PUBDEV-8147] - Upgraded XGBoost predictor to 0.3.18.
- 📦 [PUBDEV-8145] - Increased the timeout duration on the R package jar download.
- [PUBDEV-8136] - Fixed the printing for
[PUBDEV-7842] - Stacked Ensemble will no longer ignore a column if any base model uses it. [PUBDEV-7851] - Added a user-friendly reminder that the new explainability functions require newer versions of
ggplot2in R. [PUBDEV-7948] - NullPointerException error no longer thrown when used a saved and reloaded RuleFit model. [PUBDEV-7955] - Can now extract metrics from the validation dataset with a Rulefit Model. ✅ [PUBDEV-8076] - Fixed failures from Stacked Ensemble with Multinomial GLM within tests. 🛠 [PUBDEV-8077] - Fixed AutoML error when an alpha array is used for GLM. 🛠 [PUBDEV-8079] - Fixed “Rollup not possible" stats failure in GLM. [PUBDEV-8097] - H2O will now still start despite system properties that begin with ‘ai.h2o.’. 🌲 [PUBDEV-8098] - H2O exits without logging any buffered messages instead of throwing a NullPointerException when starting H2O with an invalid argument. [PUBDEV-8100] - ModelDescriptor field in MOJO is now Serializable. 🏗 [PUBDEV-8102] - AutoML no longer crashes if model builder produces H2OIllegalArgumentException in the parameter validation phase. [PUBDEV-8106] - Weights in GLM grid search is no longer used as features. 🛠 [PUBDEV-8120] - Fixed Stacked Ensemble MOJO for cases when sub-model doesn’t have the same columns as the metalearner. [PUBDEV-8125] - Efron-method now fully deterministic in CoxPH.
📜 [PUBDEV-8087] - User now allowed to specify the escape character for parsing CSVs. [PUBDEV-8092] - Added H2O reconnection script for intermittent 401 errors to R. 🙋 [PUBDEV-8101] - Added ‘ice_root’ error documented in FAQ. [PUBDEV-8118] - Added further regularization to the GLM metalearner.
🆕 New Feature
[PUBDEV-6249] - Warning now issued against irreproducible model when early stopping is enabled but neither
score_each_iterationare defined. [PUBDEV-8023] - Encrypted files that contain CSVs can now be imported. [PUBDEV-8072] - Added guidelines for correct use of
remove_collinear_columnsfor GLM. 👍 [PUBDEV-8111] - Support added for CDP 7.2.
[PUBDEV-8067] - Added information about the
pathargument for exporting .xlsx files.
[PUBDEV-6356] - GBM histograms now ignore rows with NA responses. [PUBDEV-7606] - Variable Importances added to GLM Generic model. 🛠 [PUBDEV-7782] - Fixed the ArrayIndexOutOfBoundsException issue with GLM CV. 🐎 [PUBDEV-7825] - CoxPH performance no longer fails when a factor is used for the
event_column. [PUBDEV-7841] - Existing frame no longer overwritten when data with the same query is loaded. 🛠 [PUBDEV-7909] - Fixed how
gainis calculated in XGBFI for GBM. [PUBDEV-7934] - Improved the error messages for
save_to_hive_table. ✅ [PUBDEV-7963] - Added missing argument ’test’ for
h2o.explain_row(). 🖨 [PUBDEV-7979] - All trees now supported for XGBoost Print MOJO in Java. [PUBDEV-7987] - CoxPH
predictionno longer fails when
offset_columnis specified. [PUBDEV-7998] - Added keys for Individual Conditional Expectation (ICE) plot in H2OExplanation class. [PUBDEV-8013] -
[email protected]$parameters$xnow reports actual feature names instead of
names. [PUBDEV-8016] -
h2o.explainno longer errors when AutoML object is trained with a
fold_column. 🛠 [PUBDEV-8046] - Fixed issues with python’s explanation plots not displaying fully.
🆕 New Feature
[PUBDEV-7706] - Ignored columns that are actually used for model training are unignored and no longer prevent model training to start in Flow. [PUBDEV-7735] - Added baseline hazard function estimate to CoxPH model. 👍 [PUBDEV-7748] - Target Encoding now supports feature interactions. [PUBDEV-7805] - Added CoxPH concordance to both Flow and R/Python CoxPH summaries. [PUBDEV-7820] - Added a
topbasemodelattribute to AutoML. [PUBDEV-7831] - Added new learning curve plotting function to R/Python. [PUBDEV-7854] - Added script for estimating the memory usage of a dataset. [PUBDEV-7859] - Added fault protections to grid search allowing saving of data and parameters, model checkpointing, and auto-recovery. 👍 [PUBDEV-7884] - Added support for Java 15. 👍 [PUBDEV-7969] - Added CDP7.1 support. 🖨 [PUBDEV-7978] - Added support for XGBoost to Print MOJO as JSON. 👍 [PUBDEV-8021] - Added support for refreshing HDFS delegation tokens. ⏪ [PUBDEV-8035] - Reverted XGBoost categorical encodings for contributions.
max_hit_ratio_kdeprecated and removed. 📦 [PUBDEV-7894] - Added upper bound cap to supported Java version in H2O CRAN package requirements.
[PUBDEV-7473] - Users now allowed to include categorical column name in beta constraints. [PUBDEV-7579] - Multinomial PDP can now be plotted for more than one target class in Flow. [PUBDEV-7736] - Sped up CoxPH concordance score by using tree instead of the direct approach. [PUBDEV-7819] - XGBoost no longer fails when specifying custom
fold_column. [PUBDEV-7843] - XGBoost CV models now built on multiple GPUs in parallel. [PUBDEV-7968] - Missing metrics added to GLM scoring history. [PUBDEV-8017] - Added validation checks for sampling rates for XGBoost for the R/Python clients. [PUBDEV-8024] -
No longer errors when trying to use a fold column where not all folds are represented. [PUBDEV-8032] - Added the
metalearner_transformoption to Stacked Ensemble. [PUBDEV-8057] - GBM main model now built in parallel to the CV models. 🚚 [PUBDEV-8060] - Removed redundant extraction weights from GBM/DRF histogram. [PUBDEV-8061] - GBM now avoids scoring the last iteration twice when early stopping is enabled. [PUBDEV-8063] - POJO predictions for XGBoost now even closer to in-H2O predictions. [PUBDEV-8064] - Double-scoring of CV models in AutoML now avoided thus speeding up AutoML. [PUBDEV-8070] - AutoML now uses fewer neurons in DL grids and has improved the metalearner for Stacked Ensemble.
[PUBDEV-7860] - Thin plate regression splines added to GAM.
[PUBDEV-7917] - Added checkpoint description to GLM. 📚 [PUBDEV-7976] - Added thin plate regression spline documentation to GAM algorithm page. [PUBDEV-7988] - Added missing parameters to XGBoost algorithm page. 🌲 [PUBDEV-7992] - Added more information about log files to User Guide.
[PUBDEV-7798] - GAM no longer creates multiple knots at the same coordinates when the cardinality of the
gam_columnsis less than the number of
knotsspecified by the user.
🔋 Feature interactions can now be save as .xlxs files. 👷 [PUBDEV-8034] - Job polling will retry connecting to h2o nodes if connection fails.
[PUBDEV-7949] - Partial Dependence Plot no longer failing for High Cardinality even when
user_splitsis defined. 🛠 [PUBDEV-7951] - Fixed failing Delta Lake import for Python API. [PUBDEV-7962] - Fix Stacked Ensemble’s incorrect handling of fold column.
👍 [PUBDEV-7737] - Added MOJO support for CoxPH. 0️⃣ [PUBDEV-7953] - Escape all quotes by default when writing CSV.
📄 [PUBDEV-7945] - Added to docs that AUCPR can be plotted. ⚡️ [PUBDEV-7964] - Updated the Customer Algorithm graphic for the Architecture section of the User Guide. ⚡️ [PUBDEV-7983] - Updated the copyright year to 2021.
[PUBDEV-7773] - The
pca_implparameter is no longer passed to PCA MOJO. 🚚 [PUBDEV-7896] - Objects to be retained no longer removed during the
h2o.removeAll()command. [PUBDEV-7902] - Starting GridSearch in a fresh cluster with new hyperparameters that overlap old ones will no longer cause the old models to be trained again. 0️⃣ [PUBDEV-7914] - GridSearch no longer hangs indefinitely when not using the default value for paralellism. 🛠 [PUBDEV-7921] - Fixed the parent dir lookup for HDFS grid imports. ✅ [PUBDEV-7928] - Fixed the CustomDistribution test error.
🆕 New Feature
[PUBDEV-5923] - Cross-Validation predictions can now be saved alongside the model. 👍 [PUBDEV-7269] - Added multinomial and grid search support for AUC/PR AUC metrics. [PUBDEV-7861] - Now offers a standalone R client that doesn’t include the h2o jar. 🐳 [PUBDEV-7871] - Created a Red Hat certification for H2O Docker Image. 🛠 [PUBDEV-7880] - Fixed randomized split points for
📜 [PUBDEV-7916] - Single quote regime for CSV parser exposed for importing & uploading files.
[PUBDEV-7753] - REST API disabled on non-leader Kubernetes nodes. 🖨 [PUBDEV-7875] - GLM now uses proper logging instead of printlines.
➕ Added non-tree-based models to the variable importance page in the user guide. ⚡️ [PUBDEV-7869] - Updated the AutoML citation in the User Guide to point to the H2O AutoML ICML AutoML workshop paper. ⚡️ [PUBDEV-7882] - Updated Python docstring examples about cross-validation. [PUBDEV-7905] - Corrected
kparameter description for PCA. [PUBDEV-7922] - Corrected the RuleFit Python example.
[PUBDEV-7793] - Implemented deserialization of monotone constraints. ⚡️ [PUBDEV-7844] - Updated required version of ggplot2 in R package to 3.3.0. 📜 [PUBDEV-7866] - Fixed the parsing of GLM’s
rand_familyparams in MOJO JSON. 🛠 [PUBDEV-7876] - Fixed NPE that resulted when starting a grid with SequentialWalker in AutoML exploitation phase. 🛠 [PUBDEV-7879] - Fixed MOJO version check message. [PUBDEV-7886] - When grid search has parallelism enabled, it now includes CV models.
🆕 New Feature
[PUBDEV-7739] - Added feature interactions and importance for XGBoost and GBM. [PUBDEV-7774] - Added new
interaction_constraintsparameter to XGBoost. [PUBDEV-7838] - Added an option to not have quotes in the header during exportFile. [PUBDEV-7887] - Added ability to retrieve a list of all the models in an H2O cluster. [PUBDEV-7900] - Added custom pod labels for HELM charts.
[PUBDEV-7835] - Added
lambda_maxparameters to GLMModelOutputs.
0️⃣ [PUBDEV-7745] - Added default values to all algorithm parameters in the User Guide. 🛠 [PUBDEV-7749] - Fixed the discrepancies between the Target Encoding User Guide page and Client. 📚 [PUBDEV-7834] - Added ONNX support to the documentation.
[PUBDEV-7846] - Added a new method which properly locks H2O Frames during conversion from Spark Data Frames to H2O Frames in Sparkling Water.
🛠 [PUBDEV-7836] - On the Grid Search User Guide page, fixed the missing syntax highlight in the Python example of the Random Grid Search section. [PUBDEV-7837] - Added
rule_generation_ntreesparameter to the RuleFit page. 📚 [PUBDEV-7877] - Added documentation for GBM and XGBoost on feature interactions and importance. [PUBDEV-7888] - Added a Python example to the
stratify_byparameter. [PUBDEV-7898] - Added a Feature Engineering section to the Data Manipulation page in the User Guide.
👀 [PUBDEV-7667] - Fixed StackedEnsemble’s retrieval of the seed parameter value. [PUBDEV-7746] - Deserialization values of MOJO ModelParameter now work when the Value Type is int. 📜 [PUBDEV-7760] - H2O no longer uses lazy-loading for sequential zip parse. ⚡️ [PUBDEV-7762] - Updated model_type argument names for Rulefit in R.
🆕 New Feature
[PUBDEV-7241] - Quantile distributions added to monotone constraints. [PUBDEV-7319] - TargetEncoder integrated into ModelBuilder. [PUBDEV-7755] - Python client no longer instructs the user to declare a root handler in library mode. [PUBDEV-7791] - Hostname used as certificate alias to lookup machine-specific certificate allowing Hadoop users to connect to Flow over HTTPS. [PUBDEV-7796] - Added the model explainability interface for H2O models and AutoML objects in both R & Python. [PUBDEV-7720] - Added the RuleFit algorithm for interpretability. [PUBDEV-7808] - Implemented a basic HELM chart.
[PUBDEV-7763] - Rulefit model added to algorithm section of UserGuide. [PUBDEV-7786] - Added an Explainability page to the User Guide outlining the new
h2o.explain_row()functions. ⚡️ [PUBDEV-7804] - Updated the AutoML User Guide page to include the new Explainability and Preprocessing sections.
👍 [PUBDEV-5932] - Added support for Python 3.7+. [PUBDEV-7717] - Exposes names of score0 output values in MOJO. [PUBDEV-7730] - Added function to plot a Precision Recall Curve. [PUBDEV-7740] - RuleFit model represented by the set of rules obtained from trees during training. 🐎 [PUBDEV-7765] - Performance improved for exporting a Frame to CSV. [PUBDEV-7769] - GPU backend allowed in XGBoost when running multinode even when
build_tree_one_nodeis enabled. ⚡️ [PUBDEV-7787] - Updated all URLs in R package to use HTTPS. ⬆️ [PUBDEV-7790] - Upgraded to XGBoost 1.2.0.
🏗 [PUBDEV-7366] - Added cross-validation to GAM allowing users to find the best alpha/lambda values when building a GAM model. 👍 [PUBDEV-7672] - Added TargetEncoder support for multiclass problems. 🚚 [PUBDEV-7743] - Added new TargetEncoder parameter that allows users to remove original features automatically. 👍 [PUBDEV-7778] - Implemented minimal support for TargetEncoding in AutoML.
⚡️ [PUBDEV-7541] - Updated the descriptions of AutoML in R & Python packages. 📚 [PUBDEV-7781] - Made the default for
categorical_encodingin XGBoost explicit in the documentation. ⚡️ [PUBDEV-7811] - Updated the import datatype section of the Python FAQ in the User Guide. [PUBDEV-7815] - Updated the default values for
max_rule_lengthon the RuleFit page of the User Guide. ⚡️ [PUBDEV-7816] - Updated the
validation_framedefinition for unsupervised algorithms in the User Guide.
v188.8.131.52May 18, 2020