Research Article
Rapid and Nondestructive On-Site Classification Method for Consumer-Grade Plastics Based on Portable NIR Spectrometer and Machine Learning
Table 4
Performance of classification models for white commercial plastic samples.
| PCA (n) | Calibration | Prediction | ACC (%) | ACC (%) | Precision (%) | Sensitivity (%) | F1-score (100) | Micro | Macro | Micro | Macro | Micro | Macro |
| SVM | MSC | 3 | 100 | 99.58 | 99.58 | 99.59 | 99.58 | 99.58 | 99.58 | 99.58 | SNV | 2 | 100 | 93.44 | 93.44 | 95.31 | 93.44 | 93.44 | 93.44 | 93.16 | D1 | 3 | 99.89 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | SG | 3 | 99.28 | 99.58 | 99.58 | 99.59 | 99.58 | 99.58 | 99.58 | 99.58 |
| KNN | MSC | 3 | 100 | 99.58 | 99.58 | 99.59 | 99.58 | 99.58 | 99.58 | 99.58 | SNV | 2 | 100 | 99.95 | 99.95 | 99.96 | 99.95 | 99.95 | 99.95 | 99.95 | D1 | 3 | 100 | 99.58 | 99.58 | 99.59 | 99.58 | 99.58 | 99.58 | 99.58 | SG | 3 | 100 | 99.58 | 99.58 | 99.59 | 99.58 | 99.58 | 99.58 | 99.58 |
| BPNN | MSC | 2 | 100 | 98.80 | 98.80 | 98.89 | 98.80 | 98.80 | 98.80 | 98.80 | SNV | 2 | 91.52 | 97.12 | 97.12 | 97.60 | 97.12 | 97.12 | 97.122 | 97.15 | D1 | 2 | 99.83 | 99.94 | 99.94 | 99.94 | 99.94 | 99.94 | 99.94 | 99.94 | SG | 2 | 99.77 | 99.17 | 99.17 | 99.22 | 99.17 | 99.17 | 99.17 | 99.17 |
|
|