Credit-Fraud-Detection-using-Autoencoder-and-keras

The credit data set was downloaded from Kaggle and using that dataset I have trained an Autoendocer to predict which cases will be fradulent and non-fradulent. Mainly the autoencodes tries to understand the distribution of Fradulent and non-fradulent cases. It learns by learning to differentiate between the distribution between the 2 classes of transactions. So the differences in the distribution of Fradulent and Non-fradulent cases is visualized below.