Neural Network Simulation and LSTM-based Study of Expected Return

Abstract

In order to make more accurate predictions about the gold and bitcoin markets so that investors can make the best decisions, we have designed an automated buy and sell trading model. With this model, investors can make buying and selling decisions based on price trends over a period of time, thus maximizing their returns. In this paper, we build a price prediction model based on LSTM network that can predict future prices based on prices over a period of time. The prediction results and confidence level of the network can provide accurate data for the following investment decisions. In this paper, we establish an investment decision model based on price prediction. A clear judgment of the price trend over a period of time in the future allocates the existing assets according to the level of increase or decrease. Our proposed model is capable of making price forecasts with high accuracy and maximizing investors' returns on the basis of the forecasts.