Radzicki, Michael J.
This project developed a replicable process to associate stocks into clusters based on time series data, and selects an appropriate automated trading strategy for each cluster for use in trading. This process included an exploration of data pre-processing methods, selection of a clustering algorithm suited to this application, identification of an optimal investment strategy for each cluster, and the application of strategies on the algorithmically generated portfolio. Efficacy was determined through empirical comparison of gains seen in each test, with the goal of beating the market, or generating percentage greater than the change observed in the S&P 500. This process serves as a basis for future research and development in the field of applied data mining within the financial domain.
Worcester Polytechnic Institute
Electrical and Computer Engineering
Major Qualifying Project
All authors have granted to WPI a nonexclusive royalty-free license to distribute copies of the work, subject to other agreements. Copyright is held by the author or authors, with all rights reserved, unless otherwise noted.
Social Science and Policy Studies