[Retracted] Beat Wash-Sale Tax with Multigraph Convolutional Neural Networks Based Trading Strategy

Table 13

Summary for trading with adjusted probability threshold and constrained volatility.

Evenly distributed weight
Model	MDD	Return (%)	Counts	SR	Win rate (%)	Loss (10^6)	Profit (10^6)	Net volume (10^6)
Baseline	0.32	81.13	1	2.14			1.62	1.33
Random Forest	0.08	29.37	1398	2.49	0.33	1.09	0.59	0.00
XGBoost	0.07	37.93	1422	3.10	0.33	1.12	0.76	0.10
GRU	0.08	27.48	1187	2.55	0.33	1.38	0.55	−0.13
LSTM	0.04	27.34	680	3.65	0.29	1.01	0.55	0.01
Multi-GCN-GRU	0.02	104.69	1097	12.13	0.43	0.61	2.09	1.13
Multi-GCN-LSTM	0.05	107.29	960	10.61	0.45	0.83	2.15	1.08