A Hybrid Model for Financial Portfolio Optimization Based on LS-SVM and a Clustering Algorithm

Abstract

An investment decision is one of the most important financial decisions. With the aim of optimizing investment in securities from the aspect of return and risk, investors usually diversify their portfolio securities. This paper presents a hybrid model for portfolio optimization, which consist of two steps. The first step predicts future returns on the shares, and the second step, by applying hierarchical clustering algorithm, identifies various groups of shares. The test results indicate that the suggested model is suitable for optimization of a financial portfolio as a hybrid model based on selected shares, which if included in the portfolio, enable the diversification of risk.