Skip to content

Latest commit

 

History

History
20 lines (13 loc) · 1.13 KB

about.md

File metadata and controls

20 lines (13 loc) · 1.13 KB

App Goal: Wave application for H2O AutoML

Target Audience: Data Scientists

Actively Being Maintained: Yes

Last Updated: May 1, 2023

Comes with Demo Mode (pre-loaded data, models, results, etc.): Yes

Allows for Uploading and Using New Data: Yes

Detailed Description: This application allows the user to upload a CSV file and then run H2O AutoML to train as many models as the user specifies. After training completes, models are ranked on the AutoML Leaderboard, and can be individually explained using the 'Model Explain' tab. The models can be compared using a Pareto Front plot and correlation plot under the 'AutoML Viz' tab. Variables can be explained using a Partial Dependence plot.

Features:

  • AutoML Training: Allows a user to train many models using H2O AutoML using their own train/test datasets.
  • Leaderboard: View the AutoML leaderboard to rank models.
  • AutoML Viz: Shows feature importance, Shapley contributions.
  • Model Explain: Explain any model using feature importance, Shapley values, and learning curves.
  • Deployment: Download any model in the MOJO model format.