search

bargavj / EvaluatingDPML

This project's goal is to evaluate the privacy leakage of differentially private machine learning models.
MIT License
129 stars 49 forks source link

purchase100 #1

Closed oi02lyl closed 5 years ago

oi02lyl commented 5 years ago

Could you also share the parameters (learning rate, regularization parameter, etc) for purchase100?

And are you using the dataset from here ?

bargavj commented 5 years ago

The default hyperparameter values are mentioned in the main function of attack.py. Learning Rate = 0.01, Regularization parameter = 10^-8 for neural network and 10^-5 for logistic regression (also mentioned in Readme file).

I downloaded the data set from Kaggle site (https://www.kaggle.com/c/acquire-valued-shoppers-challenge/data) and preprocessed it to get the data set used in the experiments. Run the preprocess_purchase.py to get the required data set files.