Wrap up

Dimensionality Reduction in R

Matt Pickard

Owner, Pickard Predictives, LLC

Chapter 1 - Dimensionality reduction, feature information

• Information - missing values, low variance, and correlation
• Information gain and feature importance
• Curse of dimensionality

Chapter 2 - Unsupervised feature selection

• Feature selection vs. feature extraction
• Unsupervised feature selection:
  • missing value ratio filter
  • low-variance filter
  • correlation filter
• tidymodels recipe steps

Feature Selection

Feature Extraction

Chapter 3 - Supervised feature selection

• Reviewed model building with tidymodels
• Supervised feature selection methods: lasso regression, random forest
• Evaluated reduced model performance

Chapter 4 - Feature extraction

• Principal components and feature vectors
• Principal component analysis
• t-SNE
• UMAP

Congratulations!

Dimensionality Reduction in R