• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

Machine Learning Thesis Topic

boner

Joined
10/17/15
Messages
7
Points
13
Hi, quantnet, I am currently a junior in computer science at a supposedly good school. I am interested in being a quantitative developer and just had a couple of questions about relevance of my coursework to the industry. I assume algorithms is one of the best classes in any curriculum for being a quant, but I am also doing research in machine learning and wanted to know which of the principal research area, if any, have relevance with respect to algo trading / quantitative finance. I most likely have to do a senior thesis and would rather do it on something that carries some amount of weight. Thanks a lot!
 
On the software technology side, programing all that stuff in CUDA/GPU is probably a good idea.
Maybe have a look at what's on the Nvidia site in deep machine learning.

I assume algorithms is one of the best classes in any curriculum for being a quant,

I think quants have a range of skills.
 
Last edited:
Nvidia site in deep machine learning is (predominantly) on image recognition.
In this domain the deep learning (i.e. convolutional neural networks with many hidden layers) really kicks ass.
In financial area the deep learning seems to be more hype that a breakthru technology (IMO).
For instance, have a look at this presentation (and try to grasp what's wrong here):
Fraud Detection with Deep Learning at Paypal
 
I don't know much about coding for deep learning, but I assume your model will be based around your data. So don't spend hours coding a model around bad data. You can make a really simple learning model in R to test your data quickly for viability and then develop a more complex model for your data after. But then again, I'm only guessing regarding your approach.
 
Thanks for the great answers! So with respect to coming up with models, is there any reading that would help in that respect? I have a lot of experience with machine learning but not as much with quantitative finance, especially with its applications in machine learning.
 
Back
Top