A Static-Dynamic Hypergraph Neural Network Framework Based on Residual Learning for Stock Recommendation

Table 3

The results of ablation study.

Model	CSI 100						CSI 300
	IC (↑)	Rank IC (↑)	Precision@N (↑)				IC (↑)	Rank IC (↑)	Precision@N (↑)
			Top3	Top5	Top10	Top30			Top3	Top5	Top10	Top30
GRU + Attn	0.106	0.100	60.02	59.31	58.36	55.25	0.114	0.108	60.51	59.72	58.67	57.52
GRU + Attn + Sta	0.115	0.108	60.46	60.30	58.75	55.49	0.116	0.111	60.88	59.79	58.90	57.62
GRU + Attn + Sta + Dy	0.120	0.113	60.82	60.99	59.01	55.78	0.122	0.117	61.67	60.89	60.16	58.13
SD-RL	0.124	0.118	62.86	62.07	60.05	56.28	0.131	0.125	62.39	61.70	60.84	58.83

The best result in terms of each metric is indicated in bold.