方法对比

	正则化在线学习	正则化线性回归
领域	机器学习	机器学习
方法族	Machine learning	Machine learning
起源年份	2007–2013	1970–2005
提出者	Xiao, L.; Shalev-Shwartz, S.; McMahan, H. B. et al.	Hoerl & Kennard (Ridge, 1970); Tibshirani (Lasso, 1996); Zou & Hastie (Elastic Net, 2005)
类型	Online optimization framework with regularization	Penalized linear model
开创性文献	Xiao, L. (2010). Dual Averaging Methods for Regularized Stochastic and Online Optimization. Journal of Machine Learning Research, 11, 2543–2596.	Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B, 58(1), 267–288.
别名	FTRL, Follow-the-Regularized-Leader, online regularized optimization, regularized dual averaging	Ridge regression, Lasso regression, Elastic Net regression, penalized regression
相关	6	4
摘要	Regularized online learning extends the online learning paradigm by incorporating a regularization penalty into each weight update, controlling model complexity while processing data one example at a time. Algorithms such as Follow-the-Regularized-Leader (FTRL) and Regularized Dual Averaging (RDA) make this approach practical at scale, enabling sparse, well-calibrated models on streaming data.	Regularized linear regression adds a penalty term to the ordinary least-squares objective, shrinking or zeroing out coefficients to reduce overfitting and handle multicollinearity. The three main variants — Ridge (L2 penalty), Lasso (L1 penalty), and Elastic Net (combined L1+L2) — make linear regression usable even when features outnumber observations or predictors are highly correlated.