Machine Learning Techniques for Pricing, Hedging and Statistical Arbitrage in Finance
- Publication Type:
- Thesis
- Issue Date:
- 2022
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
This thesis evaluates the performance of deep learning artificial neural network (ANN) (multi-layer perceptron (MLP) and long short-term memory (LSTM)) models and parametric (Black-Scholes-Merton, Heston, Heston jump diffusion and finite moment log stable) models in the daily prediction of Standard and Poor's (S&P) 500 call option prices and delta. The first study presents an extensive empirical assessment of the forecasting performance of daily S&P 500 call option prices/moneyness and demonstrates that ANN models tend to outperform parametric models. The second study assesses the forecasting performance of daily S&P500 call option delta and the corresponding replicating portfolio value providing useful insights for short-term hedging applications. The third study considers model averaging predictions of prices, deltas and replicating portfolio values (from deep learning ANN and parametric models) and provides empirical evidence of the effectiveness of model averaging techniques, which will be valuable for short-term risk management and derivatives evaluation. The thesis also introduces a new methodology for pair trading equity ETFs, formulated by effectively applying commonly used technical indicators and machine learning algorithms (decision tree and deep learning MLP models) to the spreads produced from traditional approaches that generate unique ways to enhance returns.
Please use this identifier to cite or link to this item: