andrew2719 / Association-rule-hiding

This repository has code for association rule hiding that is done by using Genetic algorithms . Used the library DEAP for those calculations and finding the better individual for the changes

Association-rule-hiding

Evaluation Notebook

To view the evaluation notebook and see the results, please refer to the evaluation.ipynb file.

Results Folder

The results folder contains the before and after analysis results. You can find it here.

Initial Transactions

The initial_transactions.csv file contains 2500 rows randomly selected from the Assignment-1_Data.csv file. These transactions were used for the analysis.

Modified Transactions

The modified_transactions.csv file was generated using genetic algorithms with DEAP. It contains the modified transactions used for the analysis.To view the GA and its code Rule_hiding.ipynb

How the GA is applied

The GA is done by considering 3 params (0,1,2,3) where 0 -> no change, 1->remove LHS part, 2->remove RHS part, 3->remove both LHS and RHS and finding fitness. by this combined_score = (1 - avg_penalty) * alpha + utility_score * beta Please refer to the evaluation notebook for a detailed analysis using these files.