This project used machine learning techniques to try and approximate the values of the VIX and VIX futures based on S\&P 500 options. A number of feedforward neural networks were trained using various network architectures and feature representations. LASSO regression was used to select a subset of the available features that appear to be more important for predictions. This subset was then used as a feature set for several neural networks. All neural networks were then compared on basis of accuracy to set what effects changes in the number of features had on the accuracy of the resulting model.

• This report represents the work of one or more WPI undergraduate students submitted to the faculty as evidence of completion of a degree requirement. WPI routinely publishes these reports on its website without editorial or peer review.

Creator

• Gubrud, Scott B.

Publisher

• Worcester Polytechnic Institute

Identifier

• 4626
• E-project-110320-151751

Advisor

• Paffenroth, Randy Clinton

Year

• 2020

Center

• Zurich, Switzerland Project Center - MQP

Sponsor

• ZHAW

Date created

• 2020-11-03

Resource type

• Major Qualifying Project

Major

• Computer Science

Rights statement

• In Copyright

Last modified

• 2021-02-01