WordNet-Similarity is a collection of Perl modules for the WordNet system. They are designed as object classes with methods that take two word senses as input and return the semantic relatedness of these word senses.
|Tags||education Scientific/Engineering Database Front-Ends Text Processing Linguistic Utilities|
Release Notes: A new module WordNet::Similarity::FrequencyCounter, containing common support code for information content programs, was created.
Release Notes: The vector was modified to look like the other measures and is now derived from WordNet::Similarity.pm. Some minor typos in Makefile.PL were fixed, a test case for vector was added to t/access.t, and a config option name conflict in WordNet::Similarity.pm was fixed.
Release Notes: A new measure of semantic relatedness was added, based on co-occurrence vectors of WordNet glosses. The package is set up so that similarity.pl and the other Perl utilities get installed in "/usr/local/bin". similarity.pl was completely rewritten with cleaner code and new functionality.
Release Notes: A new rawtextFreq.pl utility was added for generating information content files from plain text. similarity.pl now supports an option for specify the part-of-speech of input words while measuring relatedness. A configuration and information content file can be specified in similarity.pl. A Resnik counting option and an Add-1 smoothing option were added to the information content generation utilities. Various bugfixes were made.
Release Notes: A trace bug in hso.pm was fixed, and test cases were added for all modules.