Automatic and Efficient Parameter Selection in LIBLINEAR
After version 2.0,
a new function to select the regularization parameter by cross validation.
It automatically generates a sequence of regularization parameters
and outputs the best one according to the cross-validation accuracy.
The warm-start technique is applied to speed up the
Currently only primal classification solvers are supported (-s 0 or -s 2).
Technical details are in the following paper.
B.-Y. Chu, C.-H. Ho, C.-H. Tsai, C.-Y. Lin, and C.-J. Lin.
Warm Start for Parameter Selection of Linear Classifiers.
ACM KDD 2015.
If you successfully used this code for your applications, please let us
know. We are interested in how it's being used.
How to Run Parameter Selection in LIBLINEAR
It is very simple by using the new option -C. Examples:
Find parameters by five-fold cross validation and logistic regression.
> train -s 0 -C -v 5 heart_scale
Find parameters by ten-fold cross validation and L2-loss SVM.
> train -s 2 -C -v 10 heart_scale
Supplementary materials include the derivation not shown in the paper and more results of experiments.
Please contact Chih-Jen Lin for any question.