Statistical Learning for Financial Time Series

Given a time series of returns for a portfolio of financial instruments, develop a model that accurately predicts returns which maximize profits. The objective function will take an input of financial indicators from the previous time interval and the returns from the current time interval. These indicators can explain relationships between financial instruments in the portfolio of interest, thus are important for explaining their returns and associated risk. A common challenge with these types of problems is how easy it can be to over-fit your model. In this project, we seek to explore state-of-the-art machine learning models to determine a model with high prediction accuracy that generalized well to unseen data.

Cody Mazza-Anthony
Faculty Supervisor: 
Mark Coates
Partner University: