Research on Investment Portfolios Based on Machine Learning
The randomness of the stock market makes how to use known information to construct an investment portfolio and achieve maximum returns while avoiding risks an important issue of concern.The entropy weight-TOPSIS method was used to assign weights to the indicators and rank and select the CSI 300 constituent stocks.The random forest model and LSTM model were used to predict stock closing prices.The top 7 stocks in monthly returns were selected for portfolio construction using the mean-variance model.The returns of investment strategies under different prediction models were analyzed.The experiments ultimately verify that the investment portfolio selected using the LSTM prediction model achieves higher returns and has greater return stability.