- Used traditional technical analysis, GFTD timing signal and rotation strategy to design basic trading strategies
- Proposed dynamic fusion method from portfolio management science to optimize the weight allocation
- Performed robustness checking and sensitivity analysis towards transaction costs
-
Rotation strategy:
-
GFTD signal hyperparameter grid search (bitcoin only) with RoMaD metric:
-
Dynamic strategy fusion pipline:
-
Proportion of individual strategy in the fused version:
-
Cumulative return of investors with different levels of risk preferences:
-
Effect of commission fee with portfolio terminal value
-
Commission fee ratio and transaction frequency