A Neural Stochastic Volatility Model
This work addresses the problem of improving volatility estimation and prediction for financial time series analysis, which is crucial for investors and risk managers, by proposing a novel neural network-based approach.
The paper introduces a neural stochastic volatility model that integrates statistical models with deep recurrent neural networks to formulate volatility models. This model, consisting of generative and inference stochastic recurrent neural networks, demonstrates superior volatility estimation and prediction on real-world stock price datasets, outperforming mainstream methods like GARCH, stochvol, and GPVol in terms of average negative log-likelihood.
In this paper, we show that the recent integration of statistical models with deep recurrent neural networks provides a new way of formulating volatility (the degree of variation of time series) models that have been widely used in time series analysis and prediction in finance. The model comprises a pair of complementary stochastic recurrent neural networks: the generative network models the joint distribution of the stochastic volatility process; the inference network approximates the conditional distribution of the latent variables given the observables. Our focus here is on the formulation of temporal dynamics of volatility over time under a stochastic recurrent neural network framework. Experiments on real-world stock price datasets demonstrate that the proposed model generates a better volatility estimation and prediction that outperforms mainstream methods, e.g., deterministic models such as GARCH and its variants, and stochastic models namely the MCMC-based model \emph{stochvol} as well as the Gaussian process volatility model \emph{GPVol}, on average negative log-likelihood.