This package implements the unsupervised cross-domain word representation learning method proposed in the following paper.
@inproceedings{Bollegala:ACL:2015,
Author = {Danushka Bollegala and Takanori Maehara and Ken-ichi Kawarabayashi},
Booktitle = {Proc. of the Annual Conference of the Association for Computational Linguistics (ACL)
and the International Joint > Conferences on Natural Language Processing (IJCNLP)},
Title = {Unsupervised Cross-Domain Word Representation Learning},
Year = {2015}}
Please refer the paper for the details of the algorithms.
Source code can be found in /src.
The pre-processed Amazon multi-domain sentiment dataset can be found in /data