Abstract
A Machine Learning Approach to Yield Curve Forecasting
by
Rajiv Sambasivan
Accurate forecasting of bond yields is a problem of great interest in Computational Finance. In this work we explore a non-parametric machine learning approach to this problem. In particular, we present an approach based on using Gaussian Processes. Gaussian Processes have enjoyed great success when used for functional data in various application domains. Gaussian Processes can be used in conjunction with conventional techniques for yield curve forecasting like the Nelson-Siegel model or the Dynamic Nelson-Siegel model. This work reports the suitability of Gaussian Processes to yield curve forecasting, both in isolation and in conjunction with current techniques.
Committee
Workshop
Key Dates
Communication
First Conference Link