Not logged in.

Contribution Details

Type Master's Thesis
Scope Discipline-based scholarship
Title Deep Generation of Financial Time Series with GANs
Organization Unit
  • Chenxin Nie
  • Jörg Osterrieder
  • English
Institution University of Zurich
Faculty Faculty of Business, Economics and Informatics
Number of Pages 37
Date September 2021
Abstract Text The objective of this paper is to apply Wasserstein GAN with gradient penalty to financial time series and investigate the fidelity of generated financial series. Training with real daily and intra-day market data, the WGAN-GP model is able to generate high-fidelity synthetic time series. The quality of generated data varies among different asset classes due to the availability and characteristics of the data. Moreover, the frequency of the data does not affect the quality of synthetic data. The features of both daily and intra-day data can be captured by the model.
Export BibTeX