Portfolio Optimization based on LSTM Neural Network Prediction

Abstract

The research of portfolio optimization is to rationally allocate capital in an uncertain environment so as to realize the balance between returns and risks. In this paper, a prediction-based multi-period portfolio model is proposed to provide investors with a more economical and reliable resource allocation scheme. It utilizes LSTM neural network to predict the future stock prices, while the improved particle swarm optimization algorithm is used to solve the problem. Finally, the feasibility and validity of the model is verified through empirical research.