Portfolio Selection Based on BP Neural Network and Black-Litterman Model

Abstract

With the revolution of technology, the rapid development of machine learning has brought new impetus to the development of finance. Combining machine learning technology with traditional investment portfolio theoretical models can effectively reduce investment risks and increase returns. The Mean- Variance model proposed by the pioneer of contemporary portfolio theory, Markowitz, defines risk as the volatility of the rate of return. For the first time, the method of mathematical statistics is applied to the study of portfolio selection. This method makes the multi-objective optimization of returns and risks achieve the best balance effect. However, because this model is too sensitive to the input value, this article will use the Black-Litterman model and use the BP neural network to make predictions for the BL model's view matrix. The empirical results show that the improved Black-Litterman model obtains excess returns compared to the Mean-Variance model and other models, which optimizes the investment portfolio.