The Application Of Single-index Model In Portfolio Risk Prediction

Single-index model (SIM) is a kind of generalized regression model-The semiparametricmodel. The model can be reduced to one-dimensional index from multidimensional variables,thus solving dimension issue, which the multivariate nonparametric can not solve, having agood statistical properties and a wide range of applications in the financial economy andbiomedical and so on.With the continuous development of China’s economy and securities market, Corporatesecurities investment has been of great concern, and there are more and more people areinvolved in the market of securities investment.Corporate securities investment plays a vitalrole in a country’s financial sector, while The control of portfolio risk is an important contentof financial management. there are many uncertain factors on the market, how to predict andcontrol the portfolio risk, which caused investors to make the right investment decisions mustbe difficult. At present, relevant researchers have obtained some effective methods by hardwork to deal with above issue.Based on the advantage of SIM, in view of the focus issue on the corporate securitiesinvestment, this thesis proceeds from the main financial indicators of the listed companies,and uses SIM to explore the relationship between corporate portfolio risk and financial indexof the enterprises, then forecast the future investment risk, and compared with the existingLogistic regression model prediction method, which is generally accepted. The paper mainlydevoted to the following aspects:First, this paper introduces the related concepts and knowledge of portfolio risk,including the main feature of the portfolio risk, the influence factors, financial index and soon. It is helpful that introduce the main financial index which influences the securities risk forapplying the SIM. In addition, some forcasting methods about portfolio risk are introduced,and the basic characteristic, advantages and disadvantages of the each method are described.Secondly, this paper introduces the development, general situation and the latest researchof the SIM are summarized, moreover, gives the estimators about unknown parameters andunknown function in the studied model.Thirdly, an empirical analysis of portfolio risk is carried out. It is through dimensionalityreduction to predict the relationship between the financial index and portfolio risk fits the SIM.By compared predicted result with Logistic regression model show that SIM is better that than Logistic regression model in the prediction of portfolio risk, so proposed method haveapplication value in the stock market.