classical-mlfoundations
Supervised learning
Learning from labelled data — regression, classification, bias-variance tradeoff, and generalisation.
Уровни глубины
L0Intro~2ч
Understands the train/test split concept; knows linear regression and logistic regression exist.
L1Basics~15ч
Trains and evaluates linear/logistic regression and k-NN; understands overfitting and the bias-variance tradeoff.
L2Working~25ч
Implements pipelines with cross-validation, hyperparameter search; applies regularisation (L1/L2); explains model decisions.
L3Advanced~40ч
Understands PAC learning, VC dimension, sample complexity bounds; applies semi-supervised and active learning.
L4Research~80ч
Contributes to learning theory, meta-learning, or distribution-shift research.
Ресурсы
L0 — Intro
L1 — Basics
L2 — Working
Ведёт к
- ExtendsFeature engineering
- ExtendsDecision trees and ensembles
- ExtendsNeural networks
- RelatedStatistics
- ExtendsEvaluation metrics
- ExtendsLinear regression
- ExtendsLogistic regression
- ExtendsSupport vector machines
- Extendsk-Nearest neighbors
- ExtendsNaive Bayes classifier
- ExtendsGradient boosting
- RelatedCross-validation and model selection