This directory includes the data set used in the following paper:

N. Yoshinaga and M. Kitsuregawa. Polynomial to Linear: Efficient
Classification with Conjunctive Features. Proc. EMNLP. 
pp. 1542--1551. 2009.

linear_m*: models trained with log linear model
kernel_m*: models trained with support vector machines
tl.train, tl.dev: training and test examples for the models
                  (tl.dev is not used in the paper)

You can try something like:
> pecco -t 0 -f tl.train -v 1 linear_m2 tl.dev
> pecco -t 1 -r 0.1 -e tl.train -f tl.train -v 1 kernel_m3 tl.dev

The identity of fstrie is given by the event filename [-f].
Remember to 'make clean' in building fstrie with a new model.

