LSTM-MPT Based Quantitative Portfolio Decision Model

Abstract

With the rise of big data trends and computer computing power, there is a tendency to build a quantitative investment decision model that allows computers to perform price prediction and decision analysis to give the best daily investment strategy. In this regard, an LSTM-MPT decision model that combines traditional investment theory in the field of finance with neural network models within the field of machine learning is proposed. The model obtains the price prediction part made by the LSTM neural network, the decision part made by combining Markowitz mean-variance model, Monte Carlo algorithm, and LSTM prediction price curve. Additionally, a comparative analysis with the four commonly used portfolio model strategies shows that the LSTM-MPT decision model is valid and reliable for long-term investments. In today's big data era, the model makes full use of historical data and computer computing power to facilitate effective price prediction and decision analysis, providing reference value for relevant people.