- author = "Zheng Yin",
- title = "Essays on Genetic Programming Applied to Financial Modelling",
- school = "Michael Smurfit School of Business, University College Dublin",
- year = "2014",
- address = "Ireland",
- month = "13 " # jan # " 2014",
- keywords = "genetic algorithms, genetic programming, Option Pricing, Exchange-Traded Options, Asset Price/Return Prediction, FTSE 100, Co-movement, Realised Volatility Estimation and Forecasting, RV, BSM Delta Hedging, Optimal Control Hedging, 1-Jul-2004, FX",
-
URL = "
http://ncra.ucd.ie/News.shtml",
-
URL = "https://www.ucd.ie/quinn/facultyresearch/phdresearch/phdgraduates/yinzheng/",
-
URL = "
http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/Zheng_submitted_thesis_Nov_2013_GP_Applied_in_Financial_Modelling.pdf",
- size = "234 pages",
-
abstract = "This thesis empirically analyses the effect of market
condition variables on modeling realised volatility and
high frequency delta hedging strategies. The
methodology used, Genetic programming (GP), allows the
incorporation of time-varying nonlinear relationships
in empirical studies. The analysis demonstrates that
variables related to the condition of the market such
as liquidity, bid-ask spread and volatility indicators
have a dynamic impact on realised volatility
forecasting and delta hedging strategies, which are two
fundamental risk management issues in the financial
derivative markets.
In the study of realised volatility (RV) modelling, RV is constructed using five-minute futures logarithmic returns. One-day horizon RV is forecasted using lagged RVs and market condition factors as dynamic and possible nonlinear explanatory factors. The out-of-sample forecasting results from the proposed model are compared with benchmark models including GARCH, ARMA, HAR and stepwise linear regression models. The model outperforms the benchmark models when compared across a number of different diagnostic measures. The study illustrates the importance of allowing market condition factors to have a dynamic and nonlinear impact on RV when modeling RV out-of-sample.
The delta hedging study examines the portfolio of an option writer who delta hedges their exposure on a high frequency basis using futures contracts. Three different deterministic rules are used to indicate when the option writer should rebalance the hedged portfolio. Rebalancing is conducted at uniform time intervals, when the underlying asset moves by a fixed number of ticks and based on a change in the delta of the option. The main contribution of this study is to propose a re-balancing trigger based on the output from an optimal hedging strategy that rebalances the portfolio based on dynamic nonlinear factors related to the condition of the market including a number of liquidity and volatility factors. The proposed optimal hedging strategy outperforms the other deterministic hedging methods with a significantly lower risk than the other strategies.",
-
notes = "draft dated Nov 2013
Zheng_submitted_thesis_Nov_2013_GP%20Applied%20in%20Financial%20Modelling.pdf
so there are probably some corrections in the final
version.
5 September 2014
Supervisor: Conall O'Sullivan and Anthony Brabazon",
