All Articles
-
05
Precision, Recall, and the F-beta Score
Why accuracy lies on imbalanced data, and how F-beta lets you choose which error costs more.
-
04
Ensemble Methods: Bagging vs. Boosting
Why bagging reduces variance while boosting reduces bias, worked through with the underlying math.
-
03
KNN vs. K-Means: Same Letter, Different Problem
One is supervised classification, the other unsupervised clustering — the name similarity is a trap.
-
02
SVD and PCA: The Same Thing, Twice
PCA is diagonalizing the covariance matrix. SVD gets there without ever forming it.
-
01
GAMs vs. Multiple Linear Regression
Trading interpretability for flexibility: how generalized additive models relax the linearity assumption without giving up structure entirely.