Style Transfer with Time Series: Generating Synthetic Financial Data

TL;DR

This paper introduces a generative model for financial time series that produces realistic synthetic data, enabling better training of deep learning models for financial applications amidst data limitations and market complexities.

Contribution

The paper presents a novel generative model for financial time series that captures market structure better than traditional stochastic processes.

Findings

Generates realistic synthetic financial paths

Embeds underlying market structure effectively

Facilitates training of deep models with large simulated datasets

Abstract

Training deep learning models that generalize well to live deployment is a challenging problem in the financial markets. The challenge arises because of high dimensionality, limited observations, changing data distributions, and a low signal-to-noise ratio. High dimensionality can be dealt with using robust feature selection or dimensionality reduction, but limited observations often result in a model that overfits due to the large parameter space of most deep neural networks. We propose a generative model for financial time series, which allows us to train deep learning models on millions of simulated paths. We show that our generative model is able to create realistic paths that embed the underlying structure of the markets in a way stochastic processes cannot.