ML.NET CLI on steroids!

[dcostea]

Hackathon Idea

ML.NET CLI is a wonderful tool for automating model generation starting from a dataset and manually selecting a few parameters like features, target feature, feature type, and more. Starting from a raw dataset can be hard to get always good accuracy so why shouldn't we automatize the data preparation part? We think we can use correlation matrix and permutation feature importance to create a nice tool to help the developers with no experience.

Your name

• Daniel Costea
• Calin Racolta

Team name

Apex Team

Brief Description

ML.NET CLI can go to the next level by automatizing parts of data preparation using correlation matrix and permutation feature importance.

Other

Are you looking for team members?

• [ ] Yes
• [X] No

[jernejk]

Awesome idea! :)

If you want to go even further, you can use genetic algorithms for permutations! https://www.blazor.ai/

[dcostea]

@jernejk Thank you for the idea! I will check.