Scholarlink

LIBLINEAR is an open-source library for large-scale linear classification. It supports logistic regression and linear support vector machines [Journal of Machine Learning Research 9 (2008) 1871–1874]. However, only ridge ( \(\ell ^2\) -norm) and lasso ( \(\ell ^1\) -norm) regularizations are available there. In this article, we show how to add support of an elastic-net regularization to LIBLINEAR’s implementation of the logistic regression with a small effort. Experiments show whether this upgrade still surpasses alternative existing procedures: SAGA (Python/scikit-learn) and glmnet (R).

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Extension to LIBLINEAR’s Logistic Regression Supporting Elastic-Net Penalty

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Jacek Klimaszewski

摘要

LIBLINEAR is an open-source library for large-scale linear classification. It supports logistic regression and linear support vector machines [Journal of Machine Learning Research 9 (2008) 1871–1874]. However, only ridge ( \(\ell ^2\) -norm) and lasso ( \(\ell ^1\) -norm) regularizations are available there. In this article, we show how to add support of an elastic-net regularization to LIBLINEAR’s implementation of the logistic regression with a small effort. Experiments show whether this upgrade still surpasses alternative existing procedures: SAGA (Python/scikit-learn) and glmnet (R).