DiTo97 / kapra

Data Protection and Privacy (90538) final project at UniGe

KAPRA

Implementation of the KAPRA algorithm for (k, P)-anonymity with l-diversity.

Authors: F. Minutoli, M. Ghirardelli, S. Bagnato, and G. Losapio.

             ,--._,--.
           ,'  ,'   ,-`.
(`-.__    /  ,'   /
 `.   `--'        \__,--'-.
   `--/       ,-.  ______/
     (o-.     ,o- /
      `. ;        \
       |:          \
      ,'`       ,   \
     (o o ,  --'     :
      \--','.        ;
       `;;  :       /
        ;'  ;  ,' ,'
        ,','  :  '
        \ \   :
         `                                                            

Dataset

All datasets are accessible in the data folder.

• Weekly sales transactions from UCI.

• News popularity in social media platforms from UCI.

Folder structure

Usage

[*] Usage: python k_P_anonymity.py <algorithm> <k_value> <P_value> <paa_value> <l_value> <dataset>

Parameters explanation

• algorithm, the (k, P)-anonymity implementation: naive or KAPRA;
• k_value, the k-anonymity constraint value;
• P_value, the P-anonymity constraint value on pattern sub-groups;
• paa_value, the piece-wise aggregate approximation (PAA) value to control the dimensionality of PRs;
• l_value, the l-diversity constraint value.