Guidance for Thesis Ideas

[Oll]

Hi all,

I hope you’re doing well and sorry for the essay (you can skip most)!
I am in need of help to find a ‘relevant’ thesis topic for my Master’s degree, which I have to construct myself.

EXTRA READING >>> I have just started MSc Scientific Computing at a relatively good university in the UK (couldn’t afford a Computational Finance degree so I’m tailoring this towards one). Rather than a physics thesis, I have managed to convince the Program Director to allow me to complete a computational finance-based thesis, as long as it has a considerable amount of original research and Python/Matlab/C++/combination programming. Unfortunately, as this degree is run by the Physics department (which has very little exposure/knowledge of financial topics), none of the professors can help with the development of the thesis ideas and cannot be my supervisor. I have been told I need to find an ‘sponsor’ professor from the Computer Science department, which luckily I have. Unluckily, he, and the CS department apparently now have far too many CS students to provide me details of a project, which means I have to develop the ideas myself, present them to CS who will then sign off to Physics to allow me to do the project.
I have just completed an MSci Physics degree and so feel confident in my ability to revise/learn any of the maths required. In the past I have taken a few financial maths modules and this year, to list a few I am taking/auditing quantitative and computational finance, maths and stats of algo trading, stochastic processes in finance, machine learning and neural computation, numerical methods for finance… <<< FINISH EXTRA READING

I do not have the knowledge or background to come up with an original idea that would require enough work for a Master’s thesis, but not go overboard and be on the way to a PhD thesis, that would also be very useful to industry (Hedgefunds and the like). Ideally, I would like it to be fairly applied maths intensive, which could be modelled/simulated/confirmed through the coding I do.

I would be extremely grateful for any help, and will be more than happy to answer any questions.
Oll

TL;DR - I need to find and develop a Computational Finance idea for my Master’s Thesis. Do you have any ideas?

 

[Anupriya]

Hi

Though these topics don't fall under the strict guidelines of "Computational Finance", however they involve a good deal of quantitative analysis and are good algo trading models:

Project Work EPAT

See if they are of any use to you!

Good luck!

 

[Oll]

Anupriya,

Thanks for the link. I'll certainly look into the topics on there.

 

[rajanS]

You could try implementing the carmona-Webster (2013) model and perhaps extend that model to see how the spread affects the market makers optimal inventory position?

I'm working on something similar

https://www.princeton.edu/~rcarmona/download/fe/self_financing.pdf

 

[Oll]

Hi RajanS ,
Thanks for the suggestion . This looks interesting . In what way are you working on something similar ? For research as a hobby, or academia , or for work ?

 

[rajanS]

Hi RajanS ,
Thanks for the suggestion . This looks interesting . In what way are you working on something similar ? For research as a hobby, or academia , or for work ?

For work I'm working on something different ...
Similar in terms of applying the thesis to a different model. These are stochastic models of limit order book. As such, it concerns with the optimal placement problem. There are numerous such models. CW model is one of them

I'm just throwing random ideas. Obviously the research you write, you would either replicate the suggested model, extend the model, and confirm or fail to reject some hypothesis you have constructed in writing such a research.

Perhaps you could exploit some of the short comings of using that particular model and suggest an improvement.

Or use a different estimating procedure when fitttkng the model ?

This is just a starting point. Obviously you would have to carry on your own literature review of the aforthmentioned topic and come up with something original.

 

[Oll]

For work I'm working on something different ...
Similar in terms of applying the thesis to a different model. These are stochastic models of limit order book. As such, it concerns with the optimal placement problem. There are numerous such models. CW model is one of them

I'm just throwing random ideas. Obviously the research you write, you would either replicate the suggested model, extend the model, and confirm or fail to reject some hypothesis you have constructed in writing such a research.

Perhaps you could exploit some of the short comings of using that particular model and suggest an improvement.

Or use a different estimating procedure when fitttkng the model ?

This is just a starting point. Obviously you would have to carry on your own literature review of the aforthmentioned topic and come up with something original.

Unfortunately, I am being made to undertake a project in my supervisor's field of research, so cannot attempt the suggestions here.

Thanks ever so much for the help and hopefully this thread can be helpful to others in similar situations.