NAME¶

linclassif - predict labels by a linear classification rule

SYNOPSIS¶

linclassif [options] example_file model_file

DESCRIPTION¶

linclassif is a program that predicts labels by a linear classification rule.

example_file is a file with testing examples in SVM^light format, and model_file is the file which contains either a binary (two-class) rule f(x)=w'*x+w0 or a multi-class rule f(x)=W'*x. These are produced svmocas(1) and msvmocas(1), respectively.

OPTIONS¶

A summary of options is included below.

-h

Show summary of options.

-v (0|1)

Set the verbosity level (default: 1)

-e

Print the classification error computed from predicted labels and labels contained in example_file.

-o out_file

Save predictions to the file out_file.

-t (0|1)

Output type:

0 ... predicted labels (default)

1 ... discriminant values

EXAMPLES¶

Train the multi-class SVM classifier from example file fiply_trn.light, using svmocas(1) with the regularization constant C=10, verbosity switched off, and save model to svmocas.model:

 svmocas -c 10 -b 1 -v 0 riply_trn.light svmocas.model

Compute the testing error of the classifier stored in svmocas.model using testing examples from riply_tst.light and save the predicted labels to riply_tst.pred:

 linclassif -e -o riply_tst.pred riply_tst.light svmocas.model

SEE ALSO¶

svmocas(1), msvmocas(1).

AUTHORS¶

linclassif was written by Vojtech Franc <xfrancv@cmp.felk.cvut.cz> and Soeren Sonnenburg <Soeren.Sonnenburg@tu-berlin.de>.

This manual page was written by Christian Kastner <debian@kvr.at>, for the Debian project (and may be used by others).