Team # 2218931 Page 5 of 25
Figure 1: Literature Review Framework of Model
1.4 Our Work
The problem requires building a quantitative trading decision model to achieve the best trading
strategy. Our work mainly includes the following:
1) Based on the price data of gold and bitcoin, the LM-BP Neural Network Model is constructed.
Use the model’s output of the predicted price to build a Recurrent Decision Model. The two
constitute a Quantitative Trading Decision Model together, providing the best daily trading strategy.
2) We prove that our strategy is optimal by optimizing parameters and setting a control group.
3) Change the commission ratio for gold and bitcoin, and analyze the impact of transaction
costs on strategies and results.
4) Wrote a memo for traders with our models, strategies and results.
In order to avoid complicated description, intuitively reflect our work process, the flow chart is
shown in Figure 2: