SALτ : Efficiently Stopping TAR by Improving Priors Estimates

We publish in this Github repository the Python code for our paper SALτ : Efficiently Stopping TAR by Improving Priors Estimates. Our method details can be found in sld/sld_stopping.py file.

Datasets

The RCV1-v2 dataset is available via scikit-learn and it will automatically be downloaded if you don't already own it. For the CLEF dataset, we make available the tf-idf vectors and their labels at https://zenodo.org/record/7142640

Reproducing experiments

In order to reproduce the experiments in our paper, install the necessary packages listed in requirements.txt. You can then execute run_experiments.py and run_experiments_sr.py to run experiments for the RCV1-v2 and the CLEF EMED 2019 datasets respectively.
Use the --help or -h option for more information on the available arguments.