CSI 796
Machine Learning Methods for Financial Data
Spring, 2016
Instructor:
James Gentle
Reading Course
Yijun Wei
Topics and Assignments
- Read paper ``Machine Learning for Market Microstructure and High Frequency
Trading'' by Michael Kearns and Yuriy Nevmyvaka.
Write a critical review:
- Describe characteristics of high-frequency transactions.
- Describe learning methods used in this paper. Were they appropriate
for this particular type of problem? Why or why not? What else could have been
done?
(This should be 5 to 10 pages.)
- Survey available software for neural nets. Write brief descriptions of
what is available.
- What software is available for deep learning?
- Give an example of an application of deep learning to studying daily
stock price data, and use one software package or function in the study.
There are several issues you might address, for example autocorrelations,
within sector correlations, estimation of stock price betas, pricing of
options, etc.
These R
functions are useful for getting daily data.
- Design an R function for deep learning. (Write input/value specs, and describe
the algorithm.)
- After discussions with instructor about the design, write and test the R function.