Showing 1–1 of 1 results for author: Joerg, O

Wasserstein GAN: Deep Generation applied on Bitcoins financial time series

Authors: Rikli Samuel, Bigler Daniel Nico, Pfenninger Moritz, Osterrieder Joerg

Abstract: Modeling financial time series is challenging due to their high volatility and unexpected happenings on the market. Most financial models and algorithms trying to fill the lack of historical financial time series struggle to perform and are highly vulnerable to overfitting. As an alternative, we introduce in this paper a deep neural network called the WGAN-GP, a data-driven model that focuses on s… ▽ More Modeling financial time series is challenging due to their high volatility and unexpected happenings on the market. Most financial models and algorithms trying to fill the lack of historical financial time series struggle to perform and are highly vulnerable to overfitting. As an alternative, we introduce in this paper a deep neural network called the WGAN-GP, a data-driven model that focuses on sample generation. The WGAN-GP consists of a generator and discriminator function which utilize an LSTM architecture. The WGAN-GP is supposed to learn the underlying structure of the input data, which in our case, is the Bitcoin. Bitcoin is unique in its behavior; the prices fluctuate what makes guessing the price trend hardly impossible. Through adversarial training, the WGAN-GP should learn the underlying structure of the bitcoin and generate very similar samples of the bitcoin distribution. The generated synthetic time series are visually indistinguishable from the real data. But the numerical results show that the generated data were close to the real data distribution but distinguishable. The model mainly shows a stable learning behavior. However, the model has space for optimization, which could be achieved by adjusting the hyperparameters. △ Less

Submitted 13 July, 2021; originally announced July 2021.