Author(s)

Zhenhua Zhong, Zhuocheng Li

Corresponding Author

Zhenhua Zhong

ABSTRACT

In this paper, we build time-series forecasting models and goal-planning models for gold and bitcoin and provide optimal gold and bitcoin rotation investment strategies based on our approach. First, we used a line graph to compare the SVM-GARCH predictions with the actual data provided by the subjects and found that the predictions were highly consistent with the actual scene. Using four metrics, MSE, RMSE, MAE, and MAPE, we demonstrate that our combined model has higher prediction accuracy than a single model. Secondly, the sensitivity analysis of the planning model was carried out using the gold and bitcoin transaction commission rates. As the gold and bitcoin transaction commission rates increased, the transaction share under the optimal combination strategy decreased, which proved the rationality of the planning model.

KEYWORDS

Investment Strategy, Risk Appetite, SVM-GARCH, Dynamic Weights, Multi-Objective Programming