Predict the Future Price Movements of Gold and Bitcoin Based on The Long Short-Term Memory Neural Network Model

Abstract

Market traders often buy and sell volatile assets to maximize their returns, usually with Bitcoin or gold commissions. We started from November 9, 2016, to predict the future price movements of gold and bitcoin, based on the long short-term memory neural network model. First, we preprocessed the data for outliers and added the prices of gold and bitcoin to the training. At the same time, based on the price of the day, we also established a strategy model based on the Sharpe ratio and particle swarm algorithm. Then, through the sensitivity analysis, we found that as the transaction fee increases, the number of transactions of gold and Bitcoin decreases significantly, and the value decreases. On the contrary, there is the same theory, which proves that our model is perfect.