BMC Microbiology

Table 1 Comparison of different feature sets

From: Machine learning and phylogenetic analysis allow for predicting antibiotic resistance in M. tuberculosis

	Support Vector Machine (SVM)			Random Forest
	All features	30% PRPS	30% Random	All features	30% PRPS	30% Random
MCC
AMK	0.720	0.752	0.302	0.881	0.883	0.481
CAP	0.539	0.620	0.334	0.779	0.780	0.569
ETH	0.325	0.370	0.269	0.550	0.605	0.488
KAN	0.766	0.685	0.415	0.812	0.856	0.546
OFL	0.508	0.549	0.294	0.778	0.778	0.452
STR	0.602	0.613	0.477	0.782	0.801	0.650
ROCAUC
AMK	0.842	0.872	0.663	0.927	0.918	0.765
CAP	0.779	0.816	0.683	0.858	0.863	0.718
ETH	0.665	0.686	0.640	0.780	0.807	0.746
KAN	0.873	0.844	0.718	0.880	0.916	0.741
OFL	0.757	0.781	0.649	0.870	0.870	0.723
STR	0.798	0.816	0.747	0.865	0.890	0.785
Sensitivity
AMK	0.721	0.786	0.485	0.868	0.843	0.671
CAP	0.650	0.720	0.517	0.733	0.760	0.453
ETH	0.603	0.603	0.676	0.689	0.779	0.691
KAN	0.787	0.755	0.585	0.777	0.681	0.532
OFL	0.719	0.771	0.573	0.771	0.771	0.625
STR	0.687	0.741	0.642	0.752	0.815	0.601
Specificity
AMK	0.965	0.958	0.840	0.987	0.994	0.858
CAP	0.910	0.912	0.849	0.983	0.979	0.982
ETH	0.727	0.769	0.603	0.810	0.835	0.802
KAN	0.961	0.933	0.851	0.983	0.992	0.949
OFL	0.796	0.790	0.725	0.970	0.970	0.820
STR	0.910	0.891	0.853	0.977	0.965	0.969

Back to article page

ISSN: 1471-2180

Contact us

General enquiries: journalsubmissions@springernature.com