Table 5. Correction of previous false negatives and false positives by current SVM model.
Sample | Error type i | Reported Z score | D1 j | D2 | D3 | D4 | D5 | D6 | D7 | D8 | D9 | D10 | Probability of negative in SVM prediction | Probability of positive in SVM prediction |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
chr13 | ||||||||||||||
13_FP_1 | FP | 3.61 | 7.80908 | -18.4017 | 10.4249 | -15.4931 | 8.12209 | -19.1393 | 22.38411066 | 151.515 | 32 | 17 | 0.998877 | 0.00112284 |
13_FP_2 | FP | 4.78 | 7.74425 | -9.47269 | 9.7306 | -7.34031 | 8.2322 | -10.0696 | 14.70334619 | 130.36 | 25 | 17 | 0.998162 | 0.00183793 |
chr18 | ||||||||||||||
18_FN_1 | FN | 1.77 | 3.31662 | -2.6893 | 5.82699 | -0.108174 | 3.20528 | -2.59902 | 5.63756896 | 119.412 | 42 | 17 | 1.71E-007 | 1 |
18_FN_2 | FN | 1.43 | 5.54394 | -3.20759 | 8.52815 | -0.100806 | 5.79965 | -3.35554 | 8.214781802 | 149.353 | 41 | 16 | 1.00E-007 | 1 |
18_FP_1 | FP | 3.22 | 3.36731 | -10.1558 | 4.56993 | -8.8769 | 3.31148 | -9.98747 | 12.69375338 | 153.482 | 26 | 17 | 0.984857 | 0.015143 |
18_FN_3 | FN | 1.93 | 5.92865 | -4.21056 | 6.68878 | -3.41427 | 4.61288 | -3.2761 | 15.25077553 | 148.944 | 27 | 18 | 2.13E-005 | 0.999979 |
chr21 | ||||||||||||||
21_FP_1 | FP | 2.27 | 2.17527 | -10.4262 | 4.38166 | -8.02835 | 2.35229 | -11.2747 | 17.3586114 | 139.857 | 25 | 17 | 0.997866 | 0.00213356 |
21_FN_1 | FN | 1.76 | 6.7041 | -3.30991 | 8.41899 | -1.47674 | 5.21281 | -2.57365 | 13.79432935 | 150.056 | 36 | 12 | 0.00435514 | 0.995645 |
The SVM model trained in this study corrected four false negatives and four false positives previously called by one-Z-test method.
i FP means false positive and FN means false negative.
j The definitions of features D1 to D10 were given in Table 2.