Research Article
A Stock Selection Model of Image Classification Method Based on Convolutional Neural Network
Table 4
Results of CNN under different parameters.
| Training size | Sample width | Type | Benchmark | Annual return | Sharp ratio | Max drawdown |
| 36 | 3 | 1 | 45285 | 64351 | 0.0146 | 0.2176 | 36 | 3 | 2 | 45285 | 61248 | 0.0196 | 0.2420 | 36 | 3 | 3 | 45285 | 20171 | 0.0134 | 0.2188 | 24 | 3 | 1 | 29864 | 12915 | −0.0011 | 0.2437 | 24 | 3 | 2 | 29864 | 75060 | 0.0243 | 0.2225 | 24 | 3 | 3 | 29864 | 15195 | −0.0135 | 0.2548 | 36 | 6 | 1 | 45285 | 6464 | −0.0003 | 0.2171 | 36 | 6 | 2 | 45285 | 27737 | 0.0159 | 0.2060 | 36 | 6 | 3 | 45285 | 58898 | 0.0194 | 0.2041 | 24 | 6 | 1 | 29864 | −6565 | −0.0038 | 0.2429 | 24 | 6 | 2 | 29864 | −24867 | −0.0210 | 0.2394 | 24 | 6 | 3 | 29864 | 12791 | −0.0013 | 0.2159 |
|
|