Table 3 Validation set experimental results of SVM, AlexNet and MCNN.
From: Near-infrared spectroscopy combined with pattern recognition algorithms to quickly classify raisins
Number of PCA components | SVM | AlexNet | MCNN | ||||||
|---|---|---|---|---|---|---|---|---|---|
Accuracy (%) | Precision (%) | Recall (%) | Accuracy (%) | Precision (%) | Recall (%) | Accuracy (%) | Precision (%) | Recall (%) | |
5 | 92.35 | 88.57 | 91.67 | 85.29 | 81.87 | 80.19 | 80.35 | 83.91 | 83.33 |
10 | 90.29 | 90.15 | 91.82 | 92.76 | 96.54 | 81.12 | 82.47 | 86.26 | 77.78 |
15 | 96.11 | 90.76 | 91.82 | 93.55 | 95.55 | 90.24 | 88.75 | 86.11 | 85.85 |
20 | 98.42 | 95.71 | 100.00 | 97.38 | 96.34 | 91.63 | 94.28 | 95.13 | 94.29 |
25 | 94.75 | 91.76 | 91.76 | 94.34 | 92.28 | 80.05 | 90.83 | 94.28 | 90.60 |
30 | 92.28 | 90.91 | 83.33 | 89.75 | 89.86 | 88.33 | 89.57 | 89.68 | 88.38 |
35 | 92.85 | 93.33 | 100.00 | 92.85 | 87.35 | 83.13 | 89.57 | 91.75 | 79.19 |
40 | 90.31 | 85.71 | 87.42 | 88.22 | 82.24 | 86.37 | 85.23 | 81.74 | 87.22 |
