This reposity includes python scripts and ipython-notebooks necessary for conducting analyses on the LFM-BeyMS dataset, available via Zenodo https://doi.org/10.5281/zenodo.3784765. For more details, we refer to our publication in https://arxiv.org/abs/2102.12188.
To reproduce our results, the python scripts and ipython-notebooks must be executed in following order:
- Identification of BeyMS and MS.ipynb: Identifies BeyMS and MS based on mainstreaminess.
- Identification and Analysis of Track Clusters.ipynb: Clustering and analysis of tracks listened by BeyMS. Additional statistics on track clusters.
- Identification and Analysis of User groups.ipynb: Assign users in BeyMS to track clusters. Additional statistics of user groups.
- Rating Dataset Generation.ipynb: Create dataset used for recommendation experiments. Includes both, BeyMS and MS.
- Recommendations.py: Run several recommendation algorithms and evaluate them groupwise.
- Visualization of Recommendation Performance.ipynb: Visualize results of recommendation experiments.
- Python 3
- numpy
- matplotlib
- pandas
- seaborn
- ast
- sklearn
- scipy
- pycountry
- umap
- hdbscan
- surprise
- statsmodels
- Peter Müllner, Know-Center GmbH, pmuellner [AT] know [MINUS] center [DOT] at (Contact)
- Dominik Kowald, Know-Center GmBH
- Markus Schedl, JKU Linz
- Christine Bauer, JKU Linz
- Eva Zangerle, University of Innsbruck
- Elisabeth Lex, Graz University of Technology