Research On Security Investment Based On The Optimized Black-Litterman Model

Recently, with the development of our trade globalization, the capital market in China is not only affected by the domestic fundamentals, but is also impacted by the global capital market. The performance of our capital market, especially the equity market, becomes much more volatility and uncertainly than before, which means more challenge to investors. The scholar always study how to allocate asset to gain high returns at the low level of risk.Markowitz, the initiator of modern portfolio theory, thought the unsystematic risk can be reduced through diversification, and he gave us a mathematical model in 1952, which is the basis of modern portfolio theory. Then, many scholars proposed different models according to different conditions of market. One of them is Black-Litterman model, which presented in 1992. This model is based on M-V model and put views of investors into the M-V model. After that, we can get a new weight of portfolio. However, the Black-Litterman model randomly chooses the views of investors, as a result of which BL model cannot perform very well in a turbulent market.So, basing on BL model, this paper uses Radial Basis Function Neural Network to get views of investors. Then, we use GJR-GARCH-M model to estimate the volatility of investors’views. At last, we get the new weight of portfolio which can improve the return of portfolio.In the empirical part, this paper select 50 stocks of Hushen 300. Using optimized BL model to get new weight of portfolio and the new return of portfolio. Then we will compare this model with M-V model, as a result of which the new model has its advantage and can be a good choice for investors.