This code is part of the dissertation 'Synthetic Data in Machine Learning' by Anna Marek

This code is divided into 6 sections:

0) Data pre-processing specific to datasets used in this project

1) classification algorithms
 - contains code for neural networks, random forest and SVM classifier
 - contains code used to assess performance of classifiers by producing confusion matrices and precision-recall curves

 2) synthetic data generation
 - contains code used to synthesise data using GAN, cGAN, WGAN, WcGAN and tGAN

 3) data quality evaluation
 - contains code for SRA, feature importance, propensity score, histograms, scatterplots and confusion matrices

 4) performance improvement
 - contains code used to train random forest models and produce figures for control and results

 data
 -contains original datasets used in the project: Credit Card Fraud, Customer Churn and Bioresponse

 GAN_global_functions.py and global_functions.py are used by many scripts and should be placed in the main directory,
 not subfolders - eg. straight into the MSc Project directory. The way paths are set up requires these functions to
 be in this specific place