How deep is your model? Network topology selection from a model validation perspective

Journal Article

2022-01-03

Permanent link to this Item
https://doi.org/10.1186/s13362-021-00116-5
 http://hdl.handle.net/11427/36275
Files
13362_2021_Article_116.pdf (2.89 MB)
Authors
Nowaczyk, Nikolai
Kienitz, Jörg
Acar, Sarp K
Liang, Qian
Journal Title

Journal of Mathematics in Industry

Link to Journal
https://www.hindawi.com/journals/jim/
Journal ISSN
Volume Title
Publisher
Publisher
Department
African Inst. of Fin. Markets and Risk Mngnt
Faculty
Faculty of Commerce
License

http://creativecommons.org/licenses/by/4.0/

Series
Abstract
Deep learning is a powerful tool, which is becoming increasingly popular in financial modeling. However, model validation requirements such as SR 11-7 pose a significant obstacle to the deployment of neural networks in a bank’s production system. Their typically high number of (hyper-)parameters poses a particular challenge to model selection, benchmarking and documentation. We present a simple grid based method together with an open source implementation and show how this pragmatically satisfies model validation requirements. We illustrate the method by learning the option pricing formula in the Black–Scholes and the Heston model.
Description
Keywords
Neural networks
Model validation
SR 11-7
Derivatives
Risk management
Pricing

Reference:

Nowaczyk, N., Kienitz, J., Acar, S.K. & Liang, Q. 2022. How deep is your model? Network topology selection from a model validation perspective. Journal of Mathematics in Industry. 12(1):1. http://hdl.handle.net/11427/36275

Collections
Journal Articles