Skip to content

kb-anonymity model combines the k-anonymity model commonly used in the data mining and database areas with the concept of program behavior preservation.

Notifications You must be signed in to change notification settings

LoScarro/kb-anonymization

Repository files navigation

kb-anonymization - DPP project

Authors: Lorenzo La Corte & Davide Scarrà

Slides: https://docs.google.com/presentation/d/1iRyUV5yCQGUDL5d8ZYKKteP6Di4lyDiZoYNuDYwEGs0/edit?usp=sharing

Usage

for install dependecies: make install for executing main: make main dataset='value' bpl='value' for executing statistics on few datasets: make test-partial for executing statistics on all datasets: make test-full for deleting output files: make clean

Description

main.py consists in only one execution of the algorithm. statistics.py consists in multiple executions of the algorithm on selected datasets and the generation of the statistics. Statistics are gathered and saved in /results and they involve:

  • total time
  • time for every module
    • program_execution
    • k_anonymization
    • constraint_generation (which also contains the generation of the new tuple)
  • number of rows in output For the k-anonymity we use the k-anonymity library anonypy, k value is set to 2 by default in main.py and to [2,5] in statistics.py.

About

kb-anonymity model combines the k-anonymity model commonly used in the data mining and database areas with the concept of program behavior preservation.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published