Research Article
Is Deep Learning for Image Recognition Applicable to Stock Market Prediction?
Table 6
Accuracy comparison for CNNs with different steps per epoch during the test period.
| ā | Steps per epoch (train / test) | Hit ratio | Specificity | Sensitivity |
| CNN1 | 400 / 100 | 0.68 | 0.7221 | 0.6413 | 800 / 200 | 0.54 | 0.5324 | 0.5434 |
| CNN2 | 400 / 100 | 0.53 | 0.6155 | 0.46 | 800 / 200 | 0.52 | 0.6257 | 0.4118 |
| CNN3 | 400 / 100 | 0.61 | 0.9304 | 0.31 | 800 / 200 | 0.52 | 0.4824 | 0.5238 |
| CNN4 | 400 / 100 | 0.54 | 0.6220 | 0.4734 | 800 / 200 | 0.54 | 0.7112 | 0.3450 |
|
|