Skip to content

Murat-U-Saglam/Homormorphic-Encryption-Evaluation

Repository files navigation

Evaluating Homomorphic Encryption and its practical effects on implementation

Requirements

All of the required packages for this program are within requirements.txt, this can be installed via

pip -r requirements.txt

The codebase can only be run on a Linux environment

It is also heavily recommended the packages are installed within a virtual environment or they will clash with your core python packages.

The logical nature of the program is to generate the dataset with the dataset_generator.py file and then run the *ipynb files within the Data/EncryptedEvaluation and Data/EncryptedTraining. However there is already a ready dataset.

The structure of the project is as follows

LinearRegression

Contains the folder with the models

EncryptedEvaluation

Contains the folder with the Encrypted Evaluation models along with their respective data and unique requirements

EncryptedTraining

Contains the folder with the Encrypted Training models along with their respective data and unique requirements

Data

Contains the output of dataset_generator.py

*_Log

Contains the logs for the Encrypted Evaluation = EE and Encrypted Training = ET

LogParser.py

This is a script that parses the logs at EE_Log, in doing so it creates a file called findings.txt which aggregates the results.