ABSTRACT

   

In this article, we explore generative models in order to build a market generator. The underlying idea is to simulate artificial multi-dimensional financial time series, whose statistical properties are the same as those observed in the financial markets. In particular, these synthetic data must preserve the first four statistical moments (mean, standard deviation, skewness and kurtosis), the stochastic dependence between the different dimensions (copula structure) and across time (autocorrelation function). The first part of the article reviews the more relevant generative models, which are restricted Boltzmann machines, generative adversarial networks, and convolutional Wasserstein models. The second part of the article is dedicated to financial applications by considering the simulation of multi-dimensional times series and estimating the probability distribution of backtest statistics. The final objective is to develop a framework for improving the risk management of quantitative investment strategies.

To find out more, download the full paper

Authors

RC - Author - LEZMI Edmond
Head of Alternative Quant Portfolio Strategy, Amundi Investment Institute
École Nationale des Ponts et Chaussées
RC - Author - RONCALLI Thierry
PhD, Head of Quant Portfolio Strategy, Amundi Investment Institute
RC - Author - Jiali Xu
Alternative Quant Portfolio Strategy, Amundi Investment Institute