Skip to content


Repository files navigation

Robustness of Meta Matrix Factorization Against Decreasing Privacy Budgets

This repository includes python scripts and ipython-notebooks necessary for conducting experiments utilizing MetaMF and NoMetaMF in the setting of decreasing privacy budgets. The five utilized datasets, i.e., Douban [1], Hetrec-MovieLens [2], MovieLens 1M [3], Ciao [4] and Jester [5] are given in this repository. Additionally, we provide code for constructing and analyzing three user groups of these datasets with a low, medium and high number of ratings (available via Zenodo:


To reproduce our results, the ipython-notebooks must be executed in the following order:

  1. Initialize Folder Structure.ipynb: Sets up a hierarchy of folders for saving the experimental results.
  2. data/jester/Generation.ipynb: Preprocessing of the Jester dataset utilized in our studies.
  3. Identification of User Groups.ipynb: Identification of users with a low, medium or high number of ratings.
  4. Train and Evaluate Models.ipynb: Train and evaluate our models (i.e., MetaMF and NoMetaMF) on the provided datasets and user groups.
  5. Visualize Results.ipynb: Visualize results of our experiments.
  6. Test Personalization and Collaboration.ipynb: Visualize the item embeddings and weights of the rating prediction models.

Furthermore, includes the implementation of MetaMF and our extension NoMetaMF. However, does not need to be run.


  • Python 3
  • numpy
  • pandas
  • sklearn
  • torch
  • matplotlib


  • Peter Müllner, Know-Center GmbH, Graz, pmuellner [AT] know [MINUS] center [DOT] at (Contact)
  • Dominik Kowald, Know-Center GmbH, Graz
  • Elisabeth Lex, Graz University of Technology, Graz


[1] Hu, L., Sun, A., Liu, Y.: Your neighbors affect your ratings: on geographical neighborhood influence to rating prediction. In: SIGIR’14 (2014)

[2] Cantador, I., Brusilovsky, P., Kuflik, T.: Second international workshop on information heterogeneity and fusion in recommender systems (hetrec2011). In: RecSys’11(2011)

[3] Harper, F. M., Konstan, J. A.: The movielens datasets: History and context. ACM Transactions on Interactive Intelligent Systems (TIIS) 5(4), 1–19 (2015)

[4] Guo, G., Zhang, J., Thalmann, D., Yorke-Smith, N.: Etaf: An extended trust antecedents framework for trust prediction. In: ASONAM’14 (2014)

[5] Goldberg, K., Roeder, T., Gupta, D., Perkins, C.: Eigentaste: A constant time collaborative filtering algorithm. Information Retrieval 4(2), 133–151 (2001)


No description, website, or topics provided.






No releases published


No packages published