# H2O-3

## Docs

- [AutoML](https://mintlify.wiki/h2oai/h2o-3/algorithms/automl.md): H2O AutoML automatically trains and tunes a large selection of candidate models within a user-specified time or model limit, then ranks them on a leaderboard.
- [Clustering & Dimensionality Reduction](https://mintlify.wiki/h2oai/h2o-3/algorithms/clustering-dimensionality.md): H2O-3's unsupervised learning algorithms — K-Means for clustering and PCA for dimensionality reduction — plus a brief overview of Word2Vec for text embeddings.
- [Deep Learning](https://mintlify.wiki/h2oai/h2o-3/algorithms/deep-learning.md): H2O-3's Deep Learning algorithm implements a multi-layer feedforward neural network trained with stochastic gradient descent and backpropagation.
- [GBM & XGBoost](https://mintlify.wiki/h2oai/h2o-3/algorithms/gbm-xgboost.md): Gradient Boosting Machine (H2O native) and XGBoost — two of H2O-3's most powerful supervised learning algorithms for regression and classification.
- [GLM & GAM](https://mintlify.wiki/h2oai/h2o-3/algorithms/glm-gam.md): H2O-3's Generalized Linear Model (GLM) supports elastic net regularization across multiple distribution families. Generalized Additive Models (GAM) extend GLM with smooth spline functions.
- [Algorithm Overview](https://mintlify.wiki/h2oai/h2o-3/algorithms/overview.md): A guide to all supervised and unsupervised machine learning algorithms available in H2O-3, with guidance on choosing the right one for your problem.
- [Distributed Random Forest](https://mintlify.wiki/h2oai/h2o-3/algorithms/random-forest.md): H2O-3's Distributed Random Forest (DRF) builds an ensemble of classification or regression trees on random row and column subsets for robust, low-variance predictions.
- [Stacked Ensembles](https://mintlify.wiki/h2oai/h2o-3/algorithms/stacked-ensembles.md): H2O-3's Stacked Ensemble (Super Learner) trains a metalearner on cross-validated base model predictions to find the optimal combination of diverse algorithms.
- [H2OAutoML](https://mintlify.wiki/h2oai/h2o-3/api/python/automl.md): Automated machine learning in H2O. H2OAutoML trains and evaluates many models automatically and ranks them on a leaderboard.
- [Estimators](https://mintlify.wiki/h2oai/h2o-3/api/python/estimators.md): Overview of H2O estimators (model algorithms) and the common training API shared by all estimators.
- [h2o Module](https://mintlify.wiki/h2oai/h2o-3/api/python/h2o-module.md): Core functions exported from the h2o Python package for connecting to clusters, importing data, managing models, and interacting with H2O objects.
- [H2OFrame](https://mintlify.wiki/h2oai/h2o-3/api/python/h2oframe.md): The primary data structure in H2O. H2OFrame is a distributed data frame stored on the H2O cluster, analogous to a Pandas DataFrame or R data.frame.
- [Model Methods](https://mintlify.wiki/h2oai/h2o-3/api/python/model.md): Methods and properties available on all trained H2O model objects, including prediction, performance evaluation, variable importance, explainability, and model export.
- [R Estimator Functions](https://mintlify.wiki/h2oai/h2o-3/api/r/estimators.md): Reference for all H2O-3 model training functions available in the R package, including key parameters and usage examples.
- [R API Overview](https://mintlify.wiki/h2oai/h2o-3/api/r/overview.md): Connect to H2O-3, import data, train models, and run predictions from R using the h2o package.
- [REST API Endpoints](https://mintlify.wiki/h2oai/h2o-3/api/rest/endpoints.md): Complete reference for H2O-3 REST API endpoints covering Frames, Models, Jobs, ModelBuilders, AutoML, and Predictions.
- [REST API Overview](https://mintlify.wiki/h2oai/h2o-3/api/rest/overview.md): Use the H2O-3 REST API to manage frames, train models, run AutoML, and retrieve predictions from any HTTP client.
- [Architecture](https://mintlify.wiki/h2oai/h2o-3/concepts/architecture.md): How H2O-3 works as a distributed in-memory platform for large-scale machine learning.
- [Data munging](https://mintlify.wiki/h2oai/h2o-3/concepts/data-munging.md): Import, slice, filter, join, aggregate, and transform data in H2O-3.
- [Data structures](https://mintlify.wiki/h2oai/h2o-3/concepts/data-structures.md): The core in-memory data structures H2O-3 uses to store and distribute tabular data.
- [Docker & Kubernetes](https://mintlify.wiki/h2oai/h2o-3/deployment/docker-kubernetes.md): Run H2O in Docker containers and deploy multi-node H2O clusters on Kubernetes using the official Helm chart.
- [Hadoop & Spark](https://mintlify.wiki/h2oai/h2o-3/deployment/hadoop-spark.md): Run H2O on YARN/Hadoop clusters and Apache Spark with Sparkling Water, including cloud integrations for AWS EMR, Azure HDInsight, and GCP Dataproc.
- [MOJO & POJO Export](https://mintlify.wiki/h2oai/h2o-3/deployment/mojo-pojo.md): Export trained H2O models as MOJO or POJO artifacts for dependency-free Java scoring in production.
- [Productionizing H2O Models](https://mintlify.wiki/h2oai/h2o-3/deployment/productionizing.md): Approaches, design patterns, and performance guidance for deploying H2O models to production scoring environments.
- [Security](https://mintlify.wiki/h2oai/h2o-3/deployment/security.md): Configure authentication, TLS/HTTPS, and network security for H2O deployments. Covers Basic auth, LDAP, Kerberos, PAM, and firewall port requirements.
- [Installation](https://mintlify.wiki/h2oai/h2o-3/installation.md): Install H2O-3 via pip, conda, CRAN, or download the standalone jar. Covers requirements, verification, and Hadoop/Spark setup.
- [Introduction](https://mintlify.wiki/h2oai/h2o-3/introduction.md): H2O-3 is an open-source, distributed, in-memory machine learning platform for Python, R, Java, Scala, and the web.
- [Quickstart](https://mintlify.wiki/h2oai/h2o-3/quickstart.md): Train your first H2O-3 model in Python or R in under 5 minutes.
- [Cross-Validation](https://mintlify.wiki/h2oai/h2o-3/workflows/cross-validation.md): Use K-fold cross-validation in H2O-3 to estimate model performance without sacrificing a validation split, and access per-fold metrics and holdout predictions.
- [Model Explainability](https://mintlify.wiki/h2oai/h2o-3/workflows/explainability.md): Use h2o.explain() and h2o.explain_row() to generate variable importance plots, SHAP summaries, partial dependence plots, residual analysis, and model correlation heatmaps.
- [Grid Search & Hyperparameter Tuning](https://mintlify.wiki/h2oai/h2o-3/workflows/grid-search.md): Use H2O-3's cartesian and random grid search to tune hyperparameters, apply stopping criteria, and identify the best-performing model.
- [Saving & Loading Models](https://mintlify.wiki/h2oai/h2o-3/workflows/save-load.md): Save and load H2O-3 binary models and MOJOs to local disk, HDFS, S3, GCS, and other storage backends. Learn about model checkpointing and the download/upload workflow.
- [Training Models](https://mintlify.wiki/h2oai/h2o-3/workflows/training.md): Learn how to train supervised and unsupervised models in H2O-3, split datasets, configure common training parameters, and access model metrics.