Table 2 Diagnostic accuracy of different classification models
Comparison | Peaks included in the classification model (m/z) | Training set | Blind test set | ||
|---|---|---|---|---|---|
Recognition capability | Forecasting ability | Sensitivity | Specificity | ||
DLBP vs. CLBP | 6935.16, 1262.74, 9290.15, 5533.89, 4396.51, 5132.44, 2863.16, 1866.76, 6270.64, 4054.52, 8141.73, 4468.72, 7192.08, 1982.17, 2953.54 | 82.66% | 67.24% | 70% | 63.33% |
DLBP vs. LDH | 1692.71,1779.79 | 88.33% | 86.67% | 93.33% | 80% |
DLBP vs. N | 1779.28, 3159.06, 5292.71, 5754.98, 4054.22, 5247.91, 6934.55, 4527.3, 7473.04, 3883.04, 6270.34, 3449.05, 5265.25, 4395.98, 5131.5 | 100% | 74.71% | 86.67% | 83.3% |
LDH vs. CLBP | 1692.72, 1779.74, 7766, 4121.69, 2933.47 | 94.89% | 81.44% | 93.33% | 86.67% |
LDH vs. N | 1866.86, 2023.1, 3263.09, 4527.38, 6938.15, 7639.28, 5635.59, 6028.38 | 95% | 86.67% | 96.67 | 93.33% |
CLBP vs. N | 5755.18, 4964.69, 1450.69, 5629.16, 4018.94, 2790.29, 4467.25, 9184.37, 7062.96, 5293.26 | 95% | 76.32% | 73.33% | 73.33% |
DLBP + LDH vs. N | 7923.02, 2933.11, 2790.11, 6562.89, 5292.65, 1450.6, 4054.08, 5755.42 | 95.82% | 73.28% | 95% | 96.67% |
DLBP + LDH vs. CLBP | 1866.85, 1897.7, 8141.79, 5132.49, 5807.7, 5533.54, 4362.7, 3192.61 | 90.56% | 70.02% | 90% | 90% |
DLBP + CLBP + LDH vs. N | 5755.08, 7061.97, 2790.07, 5247.76, 1450.59, 8142.27 | 88.86% | 73.83% | 86.67% | 86.67% |