Research on Gold ETF Forecasting Based on LSTM.
ISPA/BDCloud/SocialCom/SustainCom(2019)
摘要
The gold Exchange Traded Funds (ETF) is a crucial financial product. In general, it is very difficult to make accurate price predictions because many factors can influence the price trend. The traditional approaches rarely consider other factors that affect the trend more besides historic data therefore the results obtained by the traditional methods are not reliable. In order to get a more reliable prediction, a novel model named Dilated Convolution Long Short-Term Memory (DCLSTM) considering more prediction factors based on Convolution Neural Network (CNN) and Long Short-Term Memory (LSTM). The new model is a deep neural network with better feature extraction capabilities and a broader experience field than traditional deep networks. Data has been collected from the Data Base: https://finance.yahoo.com. Test data items contains several factors that have the greatest impact on prices. The experiments show that DCLSTM has better prediction accuracy than traditional methods such as the Autoregressive Integrated Moving Average (ARIMA), K-Nearest Neighbor (KNN) and Back Propagation Neural Network (BPNN) models.
更多查看译文
关键词
Long Short-Term Memory,Convolution Neural Network,Deep Learning,Multivariate Time Series,Gold Exchange Traded Fund
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要