/FederatedNaiveBayesDP

Code from the paper "Federated Naive Bayes under Differential Privacy"

Primary LanguagePython

Federated Naive Bayes under Differential Privacy

This code can be used to reproduce the results from the paper: "Federated Naive Bayes under Differential Privacy".

Scripts:

  • python src/run.py to compute the accuracy of Naive Bayes classifier (standard, centralized DP, federated DP).
  • python src/measure_error.py to compute the errors in the parameter estimation when using DP.
  • python src/measure_sensitivity.py to compute the sensitivity of the queries.

Hyperparameters of the Monte Carlo simulations can be changed from src/config.py.