Domain Neighborhood
Machine Learning
The classical supervised-learning spine: models, losses, generalization, evaluation, and the experiment habits that make modern AI results trustworthy.
Recommended Route
Start here, then follow the prerequisites forward.
This sequence is ordered for learning rather than inventory: lower difficulty, fewer prerequisites, and more central concepts come first.
- 01Linear Regression & Least Squares
Linear regression fits a line or hyperplane by making residuals as small as possible in squared-error geometry.
16 mincodedemoafter Dot Product, DerivativesCheck Dot Product first if the symbols feel slippery.
- 02Bias-Variance Decomposition
Bias-variance decomposition separates generalization error into target mismatch, training-set instability, and irreducible noise.
19 mincodedemoafter Linear Regression & Least SquaresWhy this follows: Bias-Variance Decomposition uses Linear Regression & Least Squares directly.
- 03Classification Metrics, Thresholds, and Calibration
Classification metrics turn scores into decisions, expose threshold tradeoffs, and check whether probabilities mean what they claim.
19 mincodedemoafter Logistic Regression, Train/Dev/Test Splits, Cross-Validation, and LeakageWhy this follows: both pages keep the machine learning thread active.
- 04Multinomial Logistic Regression
Softmax regression turns one score per class into a normalized probability distribution and learns by moving probability mass toward the observed class.
18 mincodedemoafter Logistic Regression, Cross-EntropyWhy this follows: both pages keep the machine learning / classification thread active.
- 05PCA as Optimization and Eigenspace Projection
PCA turns a centered data cloud into orthonormal directions that keep variance and minimize squared reconstruction error.
20 mincodedemoafter Matrix Decompositions: Eigendecomposition, SVD, and Spectral Structure, Orthogonality, Projections, and Least-Squares GeometryWhy this follows: both pages keep the machine learning thread active.
- 06Train/Dev/Test Splits, Cross-Validation, and Leakage
Train/dev/test splits and cross-validation keep learning, choosing, and final evaluation separate; leakage breaks that contract.
18 mincodedemoafter Bias-Variance Decomposition, Linear Regression & Least SquaresWhy this follows: both pages keep the machine learning thread active.
- 07Logistic Regression
Logistic regression turns a linear score into a probability, then learns a classification boundary by Bernoulli likelihood.
17 mincodedemoafter Linear Regression & Least Squares, Maximum Likelihood, Cross-EntropyWhy this follows: both pages keep the machine learning thread active.
All Published Notebooks
Browse the territory.
Linear Regression & Least Squares
Linear regression fits a line or hyperplane by making residuals as small as possible in squared-error geometry.
Bias-Variance Decomposition
Bias-variance decomposition separates generalization error into target mismatch, training-set instability, and irreducible noise.
Classification Metrics, Thresholds, and Calibration
Classification metrics turn scores into decisions, expose threshold tradeoffs, and check whether probabilities mean what they claim.
Multinomial Logistic Regression
Softmax regression turns one score per class into a normalized probability distribution and learns by moving probability mass toward the observed class.
PCA as Optimization and Eigenspace Projection
PCA turns a centered data cloud into orthonormal directions that keep variance and minimize squared reconstruction error.
Train/Dev/Test Splits, Cross-Validation, and Leakage
Train/dev/test splits and cross-validation keep learning, choosing, and final evaluation separate; leakage breaks that contract.
Logistic Regression
Logistic regression turns a linear score into a probability, then learns a classification boundary by Bernoulli likelihood.
Model Selection and Hyperparameter Search
Model selection turns many candidate settings into one chosen procedure; dev/CV may choose, while test stays untouched for final evidence.
Regularization: Ridge, Lasso, and Elastic Net
Ridge, lasso, and elastic net add shape to coefficient space so models trade training fit for shrinkage, sparsity, and more stable validation behavior.