We present a novel approach for modeling the potentially nonlinear dependence of implied volatility surface (IVS) time series. This approach, based on a nonlinear functional autoregression (NFAR) model, effectively captures spatial and temporal dependence present in the data. The novelty is introducing neural networks that admit a Neural Tangent Kernel (NTK) parametrization to estimate this model.
Joint with Ying Chen and Hannah Lai
The NTK framework allows us to draw a connection between nonlinear function-on-function kernel regression and deep neural networks. The proposed functional Neural Tangent Kernel (fNTK) estimator empirically yields superior forecasting accuracy out-of-sample, leading to economically significant gains in option-based trading strategies relative to existing benchmarks. For example, short delta-neutral straddle trading, supported by fNTK forecasts for the S&P 500 index, achieves Sharpe ratios ranging from 1.30 to 1.83. This translates to a relative enhancement in trading outcomes of 90% to 660% relative to a linear model.
FinEML seminar series
The FinEML seminar series is designed to create a collaborative platform for the exchange of insights and findings within the field. We aim to foster a friendly atmosphere that encourages constructive feedback, providing an opportunity for both junior and senior researchers to share their work.
Submit research paper
Submit original research papers in the following topics, but not limited to:
- Asset Pricing
- Big Data
- Forecasting with Machine Learning
- Macro Finance
- Option Pricing
See also
No event items found.
