Follow this link to the repository of the code for the algorithm presented in the appendix of Pertsova (2011) implementation by Mason Chua. You can download either the .zip or the tar.gz files. Note that the optional minimization step is omitted in this implementation. To run this code, you will need to download and install OCaml on your system (see this OCaml page for the latest release and installation information), and ocamlgraph (a graph library used by the code).

The folder you download contains a sample demo file (demo.ml) which includes instructions on how to run the learning algorithm from the terminal. This file can be modified to specify the training text that will serve as the input to the learner. The text is the list of affixes along with feature sets. (Note that if you want to use privative unary features, you should encode them as binary features).

To run and compile the demo file, follow the instructions in the preamble of this file. The results will be displayed in the terminal window.

back