Abstract:Calculating true volatility is an essential task for option pricing and risk management. However, it is made difficult by market microstructure noise. Particle filtering has been proposed to solve this problem as it favorable statistical properties, but relies on assumptions about underlying market dynamics. Machine learning methods have also been proposed but lack interpretability, and often lag in performance. In this paper we implement the SV-PF-RNN: a hybrid neural network and particle filter architecture. Our SV-PF-RNN is designed specifically with stochastic volatility estimation in mind. We then show that it can improve on the performance of a basic particle filter.
Statistical Finance,Artificial Intelligence,Robotics,Econometrics,Computational Finance
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to predict the true volatility more accurately in the volatility estimation of financial markets, so as to improve the effectiveness of option pricing and risk management. Specifically, the paper focuses on the impact of market microstructure noise on volatility estimation and the limitations of existing methods in dealing with this problem:
1. **Traditional statistical methods** (such as ARCH/GARCH models) are effective, but they cannot fully consider exogenous variables and are difficult to capture some characteristics of volatility, such as long - term memory, jump effects and leverage effects.
2. **Machine learning methods** perform well in capturing time - series dependencies, but lack interpretability. And as functional models rather than probability models, they cannot provide a comprehensive understanding of the estimated volatility distribution, which is crucial for robust risk modeling.
3. **Particle filters** can approximate time - varying probability distributions through the Monte Carlo method, so as to better handle the randomness of volatility, but their performance depends on assumptions about the underlying market dynamics, which may not always hold in practical applications.
To solve these problems, the paper proposes a new hybrid model - SV - PF - RNN (Stochastic Volatility Particle Filter Recurrent Neural Network), aiming to combine the advantages of particle filters and recurrent neural networks to improve the accuracy and reliability of volatility estimation. This model allows it to learn non - linear noise distributions by introducing randomness into the input of the neural network, and accelerates the training process by pre - training the neural network separately. In addition, the paper also proposes a new loss function for learning stochastic models with random inputs, and compares the performance of this model with that of the standard particle filter through the generated data.