Abstract:
We introduce a new structural deep learning model purposing is to be able to learn price action trading features that subjective traders are using to make trading decisions based on the visual recent and actual price movements, rather than relying solely on technical indicators. The model combines convolutional neural networks (CNN) and long short-term memory (LSTM) to improve the trend forecasting of gold prices for better trading signals compared to traditional strategies. As the gold price is a time series data, it is appropriate to apply CNN and LSTM for forecasting. The concept of our model is that CNN could detect price action features or patterns in different locations of time series data; while, LSTM could maintain both short-term and long-term memory as a sequence along with time series data. The collaboration of their abilities could help the neural network model understand complex relationships between recent and actual price movements and trends in gold prices. Our study found that the combining of CNN and LSTM with price action trading features could significantly enhance trading performance in the long run.
