Predicting extreme stock performance using cooperative coevolution

Trifonov, Iavor Nikolaev

Student Work

Predicting extreme stock performance using cooperative coevolution

Public

We propose and investigate a novel cooperative co-evolutionary framework that evolves genetic algorithms in parallel to predict the performance of the stock market. We introduce several alternate methods to decompose the prediction problem, including a recursive specialization scheme, and conduct extensive experimentation to compare them. Experimental results revealed a tradeoff between classification accuracy and precision as a result of a tradeoff between specialization and generalization of the co-evolutionary scheme.

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