Regularization Paths for Generalized Linear Models via Coordinate Descent

• Main Authors: Jerome Friedman, Trevor Hastie, Rob Tibshirani
• Format: Article
• Language: English
• Published: Foundation for Open Access Statistics 2010-02-01
• Series: Journal of Statistical Software
• Subjects: lasso; elastic net; logistic regression; ℓ₁ penalty; regularization path; coordinate-descent
• Online Access: http://www.jstatsoft.org/v33/i01/paper
• ISSN: 1548-7660

We develop fast algorithms for estimation of generalized linear models with convex penalties. The models include linear regression, two-class logistic regression, and multi- nomial regression problems while the penalties include ℓ₁ (the lasso), ℓ₂ (ridge regression) and mixtures of the two (the elastic net). The algorithms use cyclical coordinate descent, computed along a regularization path. The methods can handle large problems and can also deal efficiently with sparse features. In comparative timings we find that the new algorithms are considerably faster than competing methods.