Date of Award
2023
Document Type
Dissertation
Degree Name
Ph.D.
Organizational Unit
Daniels College of Business
First Advisor
Jack Strauss
Second Advisor
Ryan Elmore
Third Advisor
Shahram Amini
Keywords
Forecasting international stock returns, Machine learning, Model averaging, Model uncertainty, Non-linearities, Parameter instability
Abstract
Using a multi-level ensemble design, we forecast international stock market returns with a novel high-dimensional data set of aggregated cross sectional firm-level predictors. The method includes considerations of model uncertainty, parameter instability, model density and non-linearities with machine learning, shrinkage and model averaging. We provide evidence that it is important to systematically focus on all four sources of forecast failure, shed light on the sparsity/density debate in the stock return forecasting dialogue and contribute interesting findings on the efficacy dimensionality reduction with principal components analysis and partial least squares. The robustness of the approach is demonstrated through applications in four international markets, the long range Welch & Goyal stock returns data set, and US interest rate forecasting.
Publication Statement
Copyright is held by the author. User is responsible for all copyright compliance.
Rights Holder
Erik Mekelburg
Provenance
Received from ProQuest
File Format
application/pdf
Language
en
File Size
107 pgs
Recommended Citation
Mekelburg, Erik, "Problems with Machine Learning, High-Dimensional Data and Forecasting Stock Returns" (2023). Electronic Theses and Dissertations. 2218.
https://digitalcommons.du.edu/etd/2218
Copyright date
2023
Discipline
Finance, Statistics