Portfolio based on Black-Litterman model with entropy compensation

The Black-Litterman model can be recognized in the financial markets in combination with subjective judgment of the investors and the traditional portfolio. Based on the model, the optimization method with entropy compensation was presented. AR-TGARCH was used on the basis of historical data to predict yields and volatility, as the input variable, instead of the pure sense of the analysts subjective decision. The optimal combination of asset weights were obtained by solving the nonlinear programming problem with entropy compensation. Empirical study shows that the new model achieves better return and has stronger applicability than other portfolio models.