Uploaded February 2019 | Updated September 2026, 3 days ago
slides: mlvu.github.io/lectures/22.Methodology2.annotated.pdf
course materials: mlvu.github.io
In this lecture we discuss how to prepare your data for machine learning project. In the second half we discuss PCA, a dimensionality reduction method which is both a good method for preparing your data, but also quite a powerful method to analyse your data and exposing its structure.
Note: the "clever man" who exposed the survivorship bias in the analysis of aircraft damage was Abraham Wald: en.wikipedia.org/wiki/Abraham_Wald
Errata: I've been a little sloppy in separating the eigendecomposition and the singular value decomposition. It's not important for the basic gist of the story, but if you need to understand the complete story of PCA, you might want to consult a more detailed treatment as well.
slides: mlvu.github.io/lectures/22.Methodology2.annotated.pdf
course materials: mlvu.github.io
In this lecture we discuss how to prepare your data for machine learning project. In the second half we discuss PCA, a dimensionality reduction method which is both a good method for preparing your data, but also quite a powerful method to analyse your data and exposing its structure.
Note: the "clever man" who exposed the survivorship bias in the analysis of aircraft damage was Abraham Wald: en.wikipedia.org/wiki/Abraham_Wald
Errata: I've been a little sloppy in separating the eigendecomposition and the singular value decomposition. It's not important for the basic gist of the story, but if you need to understand the complete story of PCA, you might want to consult a more detailed treatment as well.
