Baseline implementation for swiss dialect identification on the VarDial 2017 data set. Please use Python 2.X to run the script.
The script requires a recent version of the scikit-learn package. In most cases, installation is as easy as
pip install scikit-learn
But see http://scikit-learn.org/stable/install.html for more detailed instructions.
Then clone the repository to your local computer or one of our servers:
git clone https://github.com/bricksdont/swiss-dialect-identification
To train a baseline model, use
python baseline.py --train --model model_dummy.pkz --data train.csv --verbose --classifier dummy
To use a trained model to make predictions for the test samples:
python baseline.py --predict --samples test.csv --model model_dummy.pkz > sandboxSubmission_dummy.csv
For other options, use --help:
python baseline.py --help