11: Model Zoo

Objectives

• Identify examples of linear, tree, neighbors, and ensemble models
• Choose an appropriate model for a given dataset and task
• Apply preprocessing steps to prepare data for modeling

Reading

• More interactive articles written by Amazon’s Machine Learning team:
  • Linear Regression (originally by Amazon Web Services, some edits by Prof Arnold)
  • Random Forest
• Browse the sklearn documentation:
  • Linear Models: there’s much more here than we need; focus on the LinearRegression and LogisticRegression classes and the Polynomial Features transformer.
  • Ensembles. This section is in an odd order; I’d suggest skimming in this order:
    • Bagging
    • Random Forests
    • Gradient Boosting

Additional perspectival readings are on the Perusall assignment. Don’t try to read all of them, just pick one or two.

In case Perusall isn’t working, here’s the direct links:

• What the executive order means for openness in AI
• AI Causes Real Harm. Let’s Focus on That over the End-of-Humanity Hype - Scientific American
• Podcast: Biden’s executive order aims to limit the harms of AI - Marketplace
• The COMPASS article: Machine Bias — ProPublica

Optional Activities

If you didn’t last week…

• Try the Intro to ML lesson on Kaggle. You might even go on to try the Intermediate ML lesson.