/Credit-Exposure-Prediction-GRU

Credit Exposures using a GRU Neural Network

Primary LanguagePython

Credit Exposures using a GRU Neural Network

The repo intends to release code for our Master Thesis Neural Networks for Credit Risk and xVA in a Front Office Pricing Environment. The code can be used to

  • Produce augmented yield data from existing yield data
  • Generate expected positive exposure (EPE) from yield data
  • Train a GRU neural network to reproduce EPE from yield data

Setup

Clone the project and collect yield data to run the code.

Dependencies

Creators

Isabelle Frodé & Viktor Sambergs

Thanks

Special thanks to our supervisors Magnus Wiktorsson at Lund University and Shengyao Zhu at Nordea.