Cross Lingual Classification without translation or retraining

This project aims to create a sentiment classification model to be trained in one language and use it without retraining or translation for a new language

Dependencies

• Python 2/3 with NumPy/SciPy
• scikit-learn
• nltk

Datasets

Amazon review datasets:

• Book review dataset in data/amazon-data
• 2000 for training and 2000 for testing
• Rating used as labels for positve or negative sentiment

You can download the English (en) French (es) and German (de) embeddings this way:

# English MUSE embeddings
curl -o data/wiki.en.vec https://dl.fbaipublicfiles.com/arrival/vectors/wiki.multi.en.vec
# French MUSE Wikipedia embeddings
curl -o data/wiki.fr.vec https://dl.fbaipublicfiles.com/arrival/vectors/wiki.multi.fr.vec
# German MUSE Wikipedia embeddings
curl -o data/wiki.de.vec https://dl.fbaipublicfiles.com/arrival/vectors/wiki.multi.de.vec

Train and test classifier

This project includes testing all language pair to i.e En-En, En-Fr, En-De ,Fr-Fr, Fr-En, Fr-De, De-De, De-En, De-Fr:

To evaluate the results simply run:

python crosslingual-classification.py