Model Refinement

Description: Join Hayley to explore how to make small adjustments to machine learning models to achieve performance and better fit. This is a great module for beginners to establish a baseline understanding of model refinement that will prepare you for more advanced techniques later.

Module Type: Functional

Level: Beginner   If you are seeking something more advanced, check out our Global Settings or Advanced Configuration module series.

Audience: Model creators & managers

Prerequisites: "One AI Recipes", "One AI Exploratory Data Analysis (EDA) Report", & "One AI Results Summary" modules 

Run time: ~21 minutes (there are several refinement techniques discussed in this module; feel free to skip around)

Module Sections:

00:00 - Intro, Topics Covered, & Learning Outcomes

01:29 - Intro to Model Refinement

02:50 - The EDA & Results Summary Reports Inform Model Refinement

06:33 - Recipe Reconfiguration

• 06:38 - Recipe Reconfiguration: Adjust History Training Intervals
• 08:18 - Recipe Reconfiguration: Improve Core Attribute Selection
• 10:25 - Recipe Reconfiguration: Improve Generative Attribute Selection

11:46 - Create Separate Models for Different Parts of the Organization

13:56 - Adjust Global Settings

16:18 - Perform Per Column Interventions

20:49 - Conclusion & Thanks

The module transcript and slide deck are attached below as pdfs.

• Model Refinement Transcript.docx (2).pdf100 KB
• Refining ML Models_Slides.pdf1 MB