This project aims at detecting the fraud transactions among the Credit card transactions. Autoencoders in Keras is used for this.
Using the Autoencoders in the Keras in python, the frauds in the credit card transactions are detected.
In this stage, the data is loaded into the notebook and the fraudulent and non fraudulent data are visualized.
Encoder takes the smallest possible representation of the data and extracts the most prominent features. Decoder decodes the encoded data and seperate it into fraud and non-fraud cases.
The model is trained on 2000 non-fraud data observations
Here, the model loss, latent representations and TSNE plots are visualized.
Submitted by Rajasree Unnithan R (akshaiauh)
Download packets of source code on Coders Packet