A repository of NLP-related scripts that may be of interest to fellow researchers and developers.

The code is provided without warranty, meaning that we cannot guarantee that they will function properly and that we cannot be held responsible if something goes wrong. We might not always be able to provide support, but in most cases the author(s) of the script will be happy to hear your comments, feedback or questions.

Python tutorials Source code repository

Script Version Language Short description Download 3.0.3 python A script to test the significance of recall, precision and f-score differences between 2 machine learning systems with approximate randomization testing. 2.4.0 python Obtain evaluation statistics from TiMBL, Maxent and SVMLight files
Java Frog Wrapper 1.0 java A Frog Client in Java 1.0 python Simple module to easily send mails containing Python errors 1.1.1 python This script can be used to test the significance of a Pearson correlation coefficient and to test the difference between two correlation coefficients. 1.1.1 python A script to run a single TiMBL experiment making use of all CPUs in parallel. This should increase the speed.
template 1.0 python A template showing how scripts in the repository are documented. 1.0 python Script to compute information gain (IG) and gain ratio (GR) for words in documents that belong to different classes.