This is a hands-on introduction to H2O-3, the open source, in-memory, distributed, fast, and scalable machine learning and predictive modeling platform. With H2O-3, you can build machine learning models on big data and easily export them into production in an enterprise environment.
In this introduction, we show you how to start and connect to the in-memory H2O-3 cluster, load data, perform data wrangling tasks at big-data scale, build and evaluate predictive models (including GLM and GBM/XGBoost algorithms), and create then export a MOJO object for scoring in production. We finish with a quick introduction to the Automatic Machine Learning (AutoML) functionality provided by H2O-3.
This course primarily uses Python to connect to the H2O-3 Rest API for all commands and analyses, while also introducing H2O’s web interface, Flow.
By the end of this training, the attendee will be able to
- Start and connect to the H2O-3 server
- Load data into H2O-3
- Inspect data using H2O Flow
- Perform basic data munging tasks with H2O commands
- Fit one or more of GLM, GBM, and XGBoost models
- Create a MOJO object for production
- Build multiple competing models using AutoML
This course assumes
- Some familiarity with statistical or machine learning models
- A basic understanding of Python