h2o-3

Why we included this project

H2O is a distributed, in-memory machine learning platform for training models on datasets too large for a single machine. Its AutoML searches candidate models within a time budget and returns a leaderboard with the best performer, a practical way to get a strong baseline without hand-tuning a dozen algorithms. Explainability tools are built in, and trained models can be exported as POJO or MOJO artifacts for fast scoring in production. The platform works from Python, R, Scala, and Java, plus a Flow web interface, and runs standalone or on Hadoop and Spark clusters, so it suits both individual data scientists and enterprise big-data teams.