Paysim-Fraud-Detection-using-VAE-and-BNN

The project is devided into 4 parts :

1. Data pre-processing.
2. Data Augmentation using VAE.
3. Classification using BNN.
4. Results :
   1. Exploring the parameters destributions and plotting examples.
   2. Model evaluation.

P.S : ameliorations which will be made to this notebook in the near future :

• demystificaiton of each model.
• re-training the models with the full dataset (12 million data points fully balanced) insead of 16K datapoints.
• and ofc cleaning and re-structuring for public usage.

Author : Mohamed HMINI from INSEA / M2SI