nchambers/aliases

Prerequisites
-------------

This code uses maven for compilation and dependency management.


How to Run the Code
-------------

Train and Test
./runhandles.sh <train> <test> <.vec> <.counts> null


Research Code, Not Meant for Immediate Use
-------------

This code also builds context vectors for each twitter handle, resulting in large vectors for all
seen usernames. This is a very large GB-sized data file, and not part of the repository.
I am releasing this code for others to make use of the feature functions, but it will not run
out-of-box due to the reliance on these large data files that were separately computed.

Citation:
Kevin McKelvey, Peter Goutzounis, Stephen da Cruz, and Nathanael Chambers. "Aligning Entity Names with Online Aliases on Twitter". 5th International Wkshp on NLP for Social Media, Valencia, Spain. 2017.


Sample Data
-------------
usernames-devs.txt
usernames-test.txt
usernames-train.txt

These are the twitter handles and entity names mined from Twitter's public API.