This is a package that connects the Mallet machine learning toolkit with code
that implements the MIRA [1] learning algorithm for sequences and for
classification. 

INSTALLATION

All that should be required to compile the package after it is extracted is to
run ant: 

   ant 

This will create a directory called class with the class files of this package
and another called doc with the javadoc for this package.  The compilation uses
by default the 'mallet.jar' and 'mallet-deps.jar' mallet jar file included in
the 'lib/' directory.  You can override these by specifying a class path (with
the CLASSPATH environment variable): 

   ant -DCLASSPATH=path_to_mallet/lib/mallet-deps.jar:path_to_mallet/class/

To make sure that the code works without errors, there are junit tests that run
it on random data.  to run these junit tests run

   ant tests

EXAMPLE USE

The examples/ directory includes an example of a simple classification problem
and a simple sequence tagging problem.  The example code is similar to what an
imagined user of the code might want to write to use it.  To compile and run the
examples, you need to go to the examples directory and run the build_and_run.sh
script:

    cd examples
    sh build_and_run.sh