GW/LT3 DSL (varDial 2016 shared task) This directory contains the code for GW/LT3 systems for the following sumissions: 1. Task1 closed submission 2. Task2 open and closed submission (with and without external training data) 3. tweets_conversion includes functions to buckwalter the Arabic text and to clean the tweets 4. tweets_DA collect tweets based on country 5. test_B preprocessing contains the functions specific to testB (out-of-domain) released in the shared task For more info about the system please consult the paper.
azirikly/GW_LT3-VarDial16
Discriminating Similar Languages and Dialects Identification (varDial 2016_shared task)
Python