Embedded Files
Abstract
Abstract
In this work, we demonstrate the capabilities of shallow unsupervised Machine Learning methods in a financial context. To accomplish this, we first compare the expressiveness of several simple clustering techniques. These clusters are then first screened for interpretability, and secondly leveraged as labels for a multi-class XGBoost classifier. We treat the classifier's accuracy is used as a metric for the clustering method's effectiveness, borrowing from concepts like Granger Causality.
About Us
About Us
Taehwan Kim
Taehwan is a first-year graduate student at Virginia Tech studying machine learning and healthcare. He graduated from Bucknell University in 2020 with a BSE in Computer Science. His research interests are in Quantitive Finance, Time-Series Prediction, and Data Analytics.
Matt Harrington
Matt is a first-year graduate student at Virginia Tech studying Machine Learning. He graduated from Princeton University in 2019 with a BSE in Computer Science. His research interests are in Quantitative Finance, Cryptocurrencies, and Algorithmic Game Theory.
Page updated
Report abuse