Research Article
Feasibility Study on the Use of a Portable NIR Spectrometer and Multivariate Data Analysis to Discriminate and Quantify Adulteration in Fertilizer
Table 3
The influence of preprocessing technique on identification models (fertilizer B).
| Models | Evaluation | Preprocessing treatment at 5 PCs | Raw | SNV | FD | SD | Train | Test | Train | Test | Train | Test | Train | Test |
| RF | Accuracy | 97.50 | 97.50 | 97.40 | 97.40 | 97.40 | 97.50 | 98.00 | 98.00 | Error rate | 2.50 | 2.50 | 2.60 | 2.60 | 2.60 | 2.50 | 2.00 | 2.00 | Sensitivity | 98.10 | 98.40 | 98.00 | 97.50 | 92.10 | 91.50 | 98.00 | 98.00 | Specificity | 99.60 | 99.60 | 99.20 | 99.20 | 99.60 | 99.60 | 99.60 | 99.60 |
| SVM | Accuracy | 98.70 | 98.70 | 99.30 | 99.40 | 99.50 | 99.50 | 99.70 | 99.40 | Error rate | 1.30 | 1.30 | 0.70 | 0.60 | 0.50 | 0.50 | 0.30 | 0.60 | Sensitivity | 98.20 | 99.50 | 99.70 | 99.80 | 99.50 | 99.50 | 99.40 | 99.60 | Specificity | 99.70 | 99.70 | 98.60 | 97.90 | 97.60 | 96.80 | 98.20 | 97.80 |
|
|