Application Research Of Black-Litterman Model Based On Neural Network Prediction And Stochastic Optimization

With the rapid development of the national economy and trade,the quantity and quality of listed companies are growing at a relatively fast rate,and the investment capi-tal market will inevitably become an important means of allocating capital.Capital mar-kets(especially stock markets)are not only affected by their own fundamentals,but also by the same industry,other industries and even international capital markets,making asset prices show more uncertainty and volatility than ever before.This poses a higher challenge for individual or institutional investors to allocate capital.For risk averse people,how to configure their own assets and obtain high returns under the premise of controlling risks has become a topic that academics and capital market practitioners are constantly exploring and improving.Markowitz,the founder of contemporary portfolio theory,believes that diversi-fication of portfolios can effectively disperse non-systemic risks.In 1952,the author constructed a framework for modern portfolio theory through mathematical models.With the research of different countries and different market conditions,scholars from different countries have proposed corresponding models for their respective markets.The Black-Litterman configuration asset model proposed by Black and Litterman in 1992 is a very good optimization.Based on the traditional M-V model,the model adds the subjective view of investors and solves a new investment weight applied to the as-set allocation.However,in the process of choosing a certain subjective viewpoint,the earliest Black-Litterman model is too random.In the market environment with large volatility,the model shows certain defects and deficiencies.This paper studies the stock market portfolio management based on the Black-Litterman model.In view of the fact that the view matrix of the Black-Litterman model is too subjective,this paper uses the RBF neural network to predict the opinion ma-trix and improves the Black-Litterman model.In solving the portfolio optimization problem,this paper introduces a stochastic optimization method to solve the nonlinear problem of variance.The empirical test shows the modeling process in detail.The em-pirical results show that this model can obtain higher excess returns and get a higher Sharpe ratio than the traditional M-V model and BP-BL model.