Skip to content

Paarisaa/BMI500_Differential_Privacy-

Repository files navigation

BMI500_Differential_Privacy

This code provides the histogram of original data in white color and histogram of noisy data in blue color for each epsilon, and also provides a plot of error versus values of epsilon entered.

To run the code, run the following command:

python Differential_Privacy.py path_to_dataset epsilon_value(s)

Where path_to_dataset is the path to the csv file. If the the csv file is in current directory, path_to_dataset=./adult_age_gender_race_dataset.csv and epsilon_value(s) can be either a single value or multiple values seperated by ','.

For instance, if the csv file is in the current directory ,for a single value of epsilon you can run the following command:

python Differential_Privacy.py ./adult_age_gender_race_dataset.csv 0.2

And for multiple values of epsilon run:

python Differential_Privacy.py ./adult_age_gender_race_dataset.csv 0.1,0.2,0.4,0.6,0.8,1

Output of the code is a figure of original and noisy histograms for each epsilon and also a plot of error (mean of difference between original and noisy histograms over bins of width 5). It also prints the error for each epsilon entered.

Note: You should close each generated figure to see the next one!

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages