Study On The Risk Measure Method Of Portfolio Selection Based On Hybrid Entropy

In this paper we believe random uncertainty and fuzzy uncertainty exist simultaneously and hybrid uncertainties are more realistic in the securities market, so we assumed the future return is a variable which can reflect both kinds of uncertainties. This paper uses hybrid entropy to deal with the security’s risk which contains both random uncertainty and fuzzy uncertainty. It is more realistic to use hybrid entropy to deal with hybrid uncertainties than using variance.It’s very complex to consider the correlations between securities when using hybrid entropy to measure the risk of portfolios. If the correlations between securities were ignored, the optimum solutions of a linear optimization model base on hybrid entropy would only contain one or two security. This is contrary to the original intention of investors. So a new constraint function was added in the optimization model. With this constraint function, the optimum solutions of the optimization model would be various. An account case with the sample of ten securities of SSE180Index shows that a rising weight of the new constraint function vary the optimum solutions remarkably with the combined risk climbing not much.