2007-07-28 09:09:40 +02:00
|
|
|
LIBLINEAR is a linear classifier for data with millions of instances and
|
2009-10-29 12:09:53 +01:00
|
|
|
features. It supports L2-regularized classifiers (L2-loss linear SVM,
|
|
|
|
|
L1-loss linear SVM, and logistic regression), L1-regularized classifiers
|
|
|
|
|
(L2-loss linear SVM and logistic regression).
|
2007-05-27 16:05:54 +02:00
|
|
|
|
2007-07-28 09:09:40 +02:00
|
|
|
Main features of LIBLINEAR include
|
2007-05-27 16:05:54 +02:00
|
|
|
|
2008-06-03 04:29:39 +02:00
|
|
|
- Same data format as LIBSVM and similar usage
|
2009-10-29 12:09:53 +01:00
|
|
|
- One-vs-the rest and Crammer & Singer multi-class classification
|
2008-06-03 04:29:39 +02:00
|
|
|
- Cross validation for model selection
|
|
|
|
|
- Probability estimates (logistic regression only)
|
|
|
|
|
- Weights for unbalanced data
|
2007-05-27 16:05:54 +02:00
|
|
|
|
2007-07-28 09:09:40 +02:00
|
|
|
WWW: http://www.csie.ntu.edu.tw/~cjlin/liblinear/
|
