H2O AutoML Training Course

Overview

H2O AutoML is an artificial intelligence platform that automates the process of building, selecting and optimizing large numbers of machine learning models.

This instructor-led, live training (online or onsite) is aimed at data scientists who wish to use H2O AutoML to automate the process of building and selecting the best machine learning algorithm and parameters.

By the end of this training, participants will be able to:

  • Automate the machine learning workflow.
  • Automatically train and tune many machine learning models within a specified time range.
  • Train stacked ensembles to arrive at highly predictive ensemble models.

Format of the Course

  • Interactive lecture and discussion.
  • Lots of exercises and practice.
  • Hands-on implementation in a live-lab environment.

Course Customization Options

  • To request a customized training for this course, please contact us to arrange.

Requirements

  • Experience working with machine learning models.
  • Python or R programming experience.

Audience

  • Data scientists
  • Data analysts
  • Subject matter experts (domain experts)

Course Outline

Introduction

Setting up a Working Environment

Installing H2O

Anatomy of a Standard Machine Learning Workflow

  • Data-preprocessing, feature engineering, deployment, etc.

Statistical and Machine Learning Algorithms

  • Gradient boosted machines, generalized linear models, deep learning, etc.

How H2O Automates the Machine Learning Workflow

  • Binary Classification, Regression, etc.

Case Study: Predicting Product Availability

Downloading a Dataset

Building a Machine Learning Model

Specify a Training Frame

Training and Cross-Validating Different Models

Tuning the Hyperparameters

Training two Stacked Ensemble Models

Generating a Leaderboard of the Best Models

Inspecting the Ensemble Composition

Training many Deep Neural Network Models

Troubleshooting

Summary and Conclusion

Leave a Reply

Your email address will not be published. Required fields are marked *