Research
My research interests include the following:
- Predicting extreme values in markets and financial products.
- Nonparametric Bayesian methods.
- Point-process inference.
My PhD thesis was titled “Bayesian Nonparametric Hawkes Processes with Applications”. I took the standard Hawkes process, made some of the components nonparametric and applied it to various situations like; terror attacks, fx trading and soccer goals. All of the inference was done in a Bayesian manner with the nonparametric bits using a Dirichlet process. You can download my PhD here.
Software
I have built a R software package for fitting Dirichlet processes. You
can download it within R using install.packages("dirichletprocess")
or by visiting the CRAN page
dirichletprocess. Any
bugs, questions or help can be discussed on the github page.
I’ve also built a Julia package for fitting Hawkes processes. You can
download it using the package manager in Julia (add
HawkesProcesses
). If you like maths you can read the derivations
here. If you want help with
the package use the
github page.
There is also the AlphaVantage Julia package that I’ve contributed to. Use this if you want an easy way of pulling in financial market data.
I’ve also got a keen interest in sports modelling, looking at how we can predict the outcomes of various sports based on available data. I briefly ran (a now defunct) website www.illiquidodds.com where I provided odds on a red card across all of English football.