A multi-objective genetic stock portfolio mining approach with investor's requests

摘要

Since various objective functions should be considered for optimizing the portfolio, this study proposes a multi-objective genetic portfolio optimization approach with user's requests for deriving Pareto solutions. The two objective functions used in this study are return on investment and suitability of a portfolio. The suitability of a chromosome consists of a portfolio penalty and an investment capital penalty, which are used to reflect the satisfaction degrees of user's requests. Experiments on real datasets were conducted to show the merits of the proposed approach.