Machine Learning Algorithms¶
Supervised Learning¶
Classification¶
- Scikit-Learn: Classification algorithms — Logistic Regression, SVM, Decision Trees, Random Forests, Gradient Boosting, k-NN, Naive Bayes
- XGBoost documentation — gradient boosted trees, often top performer on tabular data
- LightGBM documentation — Microsoft’s fast gradient boosting framework
- CatBoost documentation — Yandex’s gradient boosting with native categorical feature support
Regression¶
- Scikit-Learn: Linear Models — OLS, Ridge, Lasso, ElasticNet, Bayesian regression
- Scikit-Learn: Ensemble methods — Random Forest, AdaBoost, Gradient Boosting for regression
Unsupervised Learning¶
Clustering¶
- https://machinelearningmastery.com/clustering-algorithms-with-python/ — 10 clustering algorithms implemented in Scikit-Learn
- Scikit-Learn: Clustering — K-Means, DBSCAN, Hierarchical, Spectral, HDBSCAN
- UMAP — dimensionality reduction often used before clustering
Dimensionality Reduction¶
- Scikit-Learn: Decomposition — PCA, NMF, ICA, Factor Analysis
- t-SNE — nonlinear dimensionality reduction for visualization
- UMAP — faster alternative to t-SNE that preserves more global structure
Anomaly Detection¶
- Scikit-Learn: Outlier Detection — Isolation Forest, Local Outlier Factor, One-Class SVM
- PyOD — comprehensive Python library for outlier detection (40+ algorithms)
Reinforcement Learning¶
- Stable Baselines3 — reliable PyTorch implementations of RL algorithms (PPO, A2C, DQN, SAC)
- Gymnasium (ex-OpenAI Gym) — standard API for RL environments
- Spinning Up in Deep RL — OpenAI’s educational resource on deep RL
Time Series¶
- Scikit-Learn: Time series split — cross-validation for temporal data
- statsmodels: Time Series — ARIMA, SARIMAX, VAR, exponential smoothing
- Prophet — Meta’s forecasting tool for business time series
- Darts — unified API for classical and deep learning forecasting models
Recommendation Systems¶
- Surprise — Python library for collaborative filtering (SVD, KNN, NMF)
- LightFM — hybrid recommendation with implicit and explicit feedback
- Implicit — fast collaborative filtering for implicit feedback datasets
Optimization & Search¶
- Optuna — hyperparameter optimization framework (Bayesian, TPE, CMA-ES)
- Scikit-Optimize — sequential model-based optimization
- DEAP — evolutionary algorithms (genetic algorithms, genetic programming)
See also¶
#machine-learning #algorithms #python
Page last modified: 2026-04-16 12:02:33