A text classification and linking assignment based on a social book search and task finished by SBS lab on 2016 with LibraryThing and Reddit posts data. This assignment was the final project assignmnet of course Text Mining in Leiden University, only study/education purpose usage is allowed.
Environment config:
conda env create -f environment.yml
pip install -r requirements.txtDuring the whole experiments period, please remember to check the directory you save your data and change the usage in scripts with correct corresponding ones.
To run the experiments, first you need to process the .xml file with xml_repro.py with the original xml data or use my processed csv files directly. Detailed information of data can be found the data folder.
Then if you want to train a doc2vec model with given training data, run exp.py.
To use non-deep learning models, run svc_mlp.py with train and test data, and run deep_exp.py for deep experiments. nnmodel.py stores the applied neural network models implemented with keras.