This thesis aims at improving the effectiveness of VaR models by combining Credibility Theories — the Limited Fluctuation Credibility Theory and the Greatest Credibility Theory (the Buhlmann Model and the Buhlmann-Straub Model) in Actuarial Science. This thesis creates the models while gives the method and procedure for solving them. So Credibility-VaR of portfolio is produced, which is just the early warning value ofrisk. At last, some advice is provided.In chapter one of this thesis, I introduced the defects which are exposed in the process of using VaR model; In chapter two, I introduced the concept and computing theorem of VaR in detail; In chapter three, I introduced the credibility theory briefly. Based on a brief review of the Credibility Theories, it ran to combining these credibility models into the classical VaR model systems. With a tentative conclusion of "Credibility VaR Model" , I finally pointed out the significance and also the defect of this thesis. |