Table 3.
List of the top 10 most significant features (compounds) identified by mean decreased accuracy applying Random Forest algorithm on the GC × GC-HRTOFMS data set. For each feature, important identification metrics are provided.
| Feature identification | CAS# | 1tR (w) | 2tR (s) | MS library match | MS library probability | MSI level |
|---|---|---|---|---|---|---|
| Butadioic acid dimethyl ester | 106-65-0 | 1410 | 0.895 | 917 | 96.8 | 2 |
| Nonanal | 124-19-6 | 1542 | 2.295 | 930 | 75.4 | 1 |
| 6-Octen-1-ol, 3,7-dimethyl-, acetate | 150-84-5 | 1982 | 2.403 | 942 | 31.7 | 3 |
| Cyclopentane, 1,1,3,3-tetramethyl- | 50,876-33-0 | 1210 | 1.524 | 760 | 9.5 | 3 |
| Decane, 2,5,9-trimethyl | 62,108-22-9 | 990 | 1.385 | 730 | 9.2 | 3 |
| 1,7-Octanediol, 3,7-dimethyl- | 107-74-4 | 1478 | 2.815 | 855 | 17.3 | 3 |
| 5,9-Undecadien-2-one, 6,10-dimethyl- | 689-67-8 | 2178 | 2.81 | 930 | 48.7 | 3 |
| Acetic acid, phenyl ester | 122-79-2 | 1478 | 0.93 | 831 | 59 | 2 |
| Diphenyl ether | 101-84-8 | 2106 | 0.655 | 823 | 56.6 | 2 |
| 1-Nonene | 124-11-8 | 982 | 1.49 | 925 | 24.3 | 3 |