Hands-On Artificial Intelligence for IoT

H2O AutoML

One of the most exciting features of H2O is AutoML, the automatic ML. It is an attempt to develop a user-friendly ML interface that can be used by non-experts. H2O AutoML automates the process of training and tuning a large selection of candidate models. Its interface is designed so that users just need to specify their dataset, input and output features, and any constraints they want on the number of total models trained, or time constraint. The rest of the work is done by AutoML itself; in the specified time constraint, it identifies the best performing models, and provides a leaderboard. It has been observed that, usually, the Stacked Ensemble model, the ensemble of all the previously trained models, occupies the top position ...