Table 8 The metrics based on the proposed SGP-DNN framework.
Year | Metric | Hybrid model for quantitative indicators | Hybrid model for quantitative indicators individually | |
|---|---|---|---|---|
Metric | Hybrid SGP-LSTM for both fundamental and technical indicators | SGP-LSTM for fundamental indicators | LSTM for Technical Indicators | |
2020 | Rank IC | − 10.78% | − 6.70% | − 9.49% |
ICIR | − 8.71 | − 3.9 | − 6.97 | |
Excess R above 300 | 13.53% | − 5.83% | 8.98% | |
Excess R above 500 | 19.13% | − 1.20% | 14.25% | |
Excess R above average | 21% | 2.39% | 18.60% | |
Sharp ratio | 3.26 | 0.34 | 3.1 | |
2021 | Rank IC | − 8.64% | − 7.25% | − 7.74% |
ICIR | − 7.03 | − 5.58 | − 5.88 | |
Excess R above 300 | 52.46% | 41.29% | 42.83% | |
Excess R above 500 | 26.71% | 17.49% | 18.61% | |
Excess R above average | 12% | 8.83% | 9.92% | |
Sharp ratio | 1.00 | 1.38 | 1.59 | |
2022 | Rank IC | − 10.34% | − 9.99% | − 8.68% |
ICIR | − 8.99 | − 7.32 | − 6.54 | |
Excess R above 300 | 26.04% | 34.59% | 25.87% | |
Excess R above 500 | 24.72% | 33.12% | 24.45% | |
Excess R above average | 18% | 17.84% | 9.97% | |
Sharp ratio | 2.95 | 3.09 | 1.61 | |
