Create a MaxEnt model.

Usage:

$ python create_model.py OPTIONS <VECTORS_FILE>

Options:

   --mallet-dir DIR

      Directory where Mallet lives, it should contain the bin subdirectory. The
      default is set by the MALLET variable below.

   --cross-validation N

      There is no cross validation by default, set this to a number to turn
      validation on, for example, with 5 you will get 5-fold cross validation.

   --trainer 'MaxEnt'|'NaiveBayes'|'C45'|'DecisionTree'|...

      One of the trainers allowed by Mallet, the Mallet API at
      http://mallet.cs.umass.edu/api/ has a list somewhere in the
      cs.mallet.classify package. The default is MaxEnt.

The vectors file is the text file created by create_vectors.py. By default, this
will use the MaxEnt classifier and perform no cross validation. Edit the TRAINER
and CROSS_VALIDATION variables to change this behaviour.

The script creates the following files:

   <VECTORS_FILE>.model
   <VECTORS_FILE>.err
   <VECTORS_FILE>.out
   <VECTORS_FILE>.vect

The first one is the one needed for running the classifier. Note that if cross
validation is on then there will be no .model file, instead there will be models
for each fold.

create_model(vectors, trainer='MaxEnt', cross_validation=None, output=False)
Create a Mallet model given a vectors file.

remove_files(vectors)
Remove all files derived from the vectors file.